Air Freight Industry – White Paper
The Supply Chain and Logistics Institute H. Milton Stewart School of Industrial and Systems Engineering Georgia Institute of Technology
April 1, 2007
This white paper is the result of work done by Jon Petersen as a Research Scholar in the Industry Studies Program of the Supply chain and Logistics Institute. Jon can be contacted at Petersen@gatech.edu for any comments or questions.
This document is intended to summarize the primary characteristics of the air freight industry today. The words “air freight” and “air cargo” are used interchangeably, which admittedly is not accurate. Of course, air freight generally refers to larger parcels, where cargo is typically defined as the sum of freight, packages, and mail. There are three goals for in this document: (1) To present a macro-overview of the state of the industry, and how the model of tomorrow will differ from that of today (2) To present a comprehensive summary of how each component along the air cargo supply chain works (3) To summarize the most prominent challenges faced by the industry, and to suggest areas for future research The industry continues to emerge and faces several challenges in order to accommodate the acceleration in the volume of airfreight. We conclude by suggesting the most critical areas to address these issues.
I. Industry Facts
A. Volume ……………………………………………. B. Determinants of Air Cargo Volume ………………... C. The Geography of Freight Movement …………….... D. Air Cargo Tomorrow ……………………………….
II. How the Process Works
4 6 8 10
A. The Players Involved ….……………………………. B. Process Flow: An Overview ………………………... C. Case Study at Singapore Changi Airport ……..……....
III. The Freight Forwarder
13 13 15
A. Classes of Forwarders ….……………………………. B. Market Facts ………………………………………... C. Case Study at BAX Global …………………..……....
IV. The Carrier
18 18 19
A. Classes of Carriers ….………………….……………. B. Major Carriers ………………………………….…... C. Airports …………………..………………………....
24 C.1 HACTL ………………………………………… 25 C.2 Busiest Airports by Volume …………………….. 25 D. Aircraft ……………………………………………… 28
D.1 Overview ……………………………………….. …… 28 D.2 Freighter Data …………………………………............. 28 D.3 The Air Freighter of Tomorrow ……………………….. 30 D.4 Containers ……………………………………………… 31
E. Fright Carrier Economics …………………………….... 32
E.1 Coordination ……………………………………............ 32 E.2 Pricing ………………………………………………..… 32 E.3 Air Cargo Revenue Management …………………..…… 35
A. Air Cargo Security Risks ….…………………………. . B. Known Shipper Programs ....………………………... C. Air Cargo Security Technology …………………..…..
39 40 41
I. Industry Facts
Given the degree of globalization and the emergence of global supply chains, the role of the air freight industry has become increasing important to the global economy. Consequently, there have been considerable changes in the industry as it continues to evolve. The growth of international air freight has outpaced that of global GDP by a factor of 2.5 since 1980. Air freight traffic constitutes less than 2 percent of all tonnage transported. However, it represents over one-third of the aggregate value of all international trade. Virtually anything can be shipped vis-à-vis air freight. Most of this tends to be high-valued goods such as pharmaceuticals, machine tools, computers and electronics, aircraft, auto parts, perishables, instruments, and medical equipment. Most of this is considered just-in-time components used for time-sensitive processing. Estimates on the nominal size of the air cargo industry range typically fluctuate around $50 billion.
According to the 2006-2007 Boeing World Air Cargo Forecast in 2005 there were 178,122 million tons of freight ton-kilometers flown globally. The growth of air cargo globally has averaged a 5.1% annual growth rate since 1995 (see Figure 1). Not surprisingly, the majority of this growth is attributable to growth in Asian markets, particularly China. This growth rate is anticipated to accelerate (see below). Recall cargo is defined to be the sum of freight (scheduled and chartered), mail, and express carriers (i.e. small packages typically handled by UPS, FedEx, and DHL). Figure 2 shows the decomposition of freight by class among cargo handled in the U.S. Freight continues to exhaust the majority of all cargo – averaging 61% of all cargo over this period, while it has grown from 57% in 1995 to 64% in 2005. Recently, there has been a slight downturn in the industry. In the first half of 2006, the growth rate has been less than half relative to its ten year average. Despite this, the long term outlook for the volume of air cargo is to keep growing at an accelerated pace. According to Boeing, the total volume of air cargo is anticipated to grow at an average annual rate of 6.1% through 2025. Among this growth, Asia is expected to lead the way. • Boeing is forecasting annual growth of all cargo in domestic China go grow at 10.6% per year until 2025 • Intra-Asian routes are anticipated to grow at 8.6% annually • Asian-North American routes are expected to grow 7.1% annually; and AsianEuropean at 6.9%. That leads to the natural question of what influences air cargo?
Volume of Global Air Cargo 1995-2005
200,000 180,000 All Carriers 160,000 140,000
air-ton kilometers, millions
U.S. Carriers Non-U.S. Carriers
120,000 100,000 80,000 60,000 40,000 20,000 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
source: Boeing 2006-2007 World Air Cargo Forecast
B. Determinants of Air Cargo Volume.
Here we examine three possible dynamics regarding on what drives air cargo. Of course, deriving causation from correlation is often spurious. So we only examine a statistical relationship. We consider the relationship between the volume of air cargo an (a) the global macroeconomy (b) the price level and (c) the price of jet fuel. Figure 3 shows the relationship between World GDP1. With the exception of the 2001 recession, there appears to be an unambiguously positive relationship between the two, which is not surprising.
normalized to current-year US $; Source: World Bank
Components of Air Cargo, U.S. Carriers 1995-2005
35,000 Freight 30,000 Mail Express Carriers 25,000
air-ton kilometers, millions
0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
source: Boeing 2006-2007 World Air Cargo Forecast
Cargo Volume and World GDP 1995-2005
global air cargo (left axis) world GDP (right axis) 160 41
100 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
trillions US $
Figure 4 shows the relationship between the volume of air cargo, the (normalized) price of jet fuel, and the (normalized) price of air transportation2. One might expect a negative relationship between the price levels and the volume of air cargo. However the figure shows this is not necessarily the case. The relationship between volume and each of the price indices shows the lack of an unambiguous relationship. We therefore conclude that, at least in the past decade, the growth of air cargo has been robust. That is, the high growth of air cargo has developed in spite of unprecedented shocks concerning jet fuel.
Growth in Cargo Volume versus Jet Fuel Prices 1995-2005
Cargo volume Price of jet fuel Price of air transportation
% change from previous year
-40 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
As a coarse approach to quantifying the relationship each of these variables have on volume, let us look at a simple OLS regression. We wish to fit the following model
Ptrans t +
where Volt denotes the volume of cargo transported (ton-kilometers) in year t, WGDPt denotes world GDP in year t, Pfuelt denotes the change in the price of jet fuel in year t
this is the producer price index of jet fuel; source: Bureau of Economic Analysis
relative to year t-1, and Ptrans t denotes the change in price of air transportation in year t over t-1. An OLS produces the following output:
Volt = 25.88 + 3.48 WGDPt + 0.10 Pfuelt
( 0.58 ) ( 2.75 )
( 0.18 )
0.33 Ptrans t +
R 2 = 0.7240
Where the OLS coefficients are reported and their corresponding t-values are below. We note from this simple model that only world GDP statistically captures air cargo volume, and as we would expect, we have a positive coefficient of world GDP. The model captures most of the variation as evidence by the goodness-of-fit metric. By dropping jet fuel inflation and air transportation prices, we have an improved reduced model:
Volt = 14.72+ 3.82 WGDPt +
( 0.69 ) ( 6.04 )
R 2 = 0.7800 So we conclude by saying that world output statistically affects volume of air cargo, at least from 1995-2005. Moreover, for every trillion dollar increase in world GDP, we would expect the volume of air cargo to increase by 3.82 billion ton-kilometers, cetris paribus.
C. The Geography of Freight Movement
Here we address the movement of air freight with respect to geography. Let us examine the anticipated freight volume growth across regions. Freight within North America 56% of all air cargo is shipped with at least one endpoint in the U.S. According to the Airforwarders Association, in 2001 31% of total global freight activity was within the U.S.; 25% was U.S.-abroad; and the remaining 44% was non-U.S. activity. Within North America, lanes traversing the US-Canadian border experience a surprising asymmetry with northbound freight volumes being more than twice of southbound freight. Northbound freight typically contains small packages, computers, and general industrial machinery. Southbound freight contains mostly telecommunications equipment, small packages, and electric machinery. North America – Asia North American-Asian routes have boomed from the coupled effect both globalization and strong domestic growth. Thanks to the two the continued economic expansion of China, it is now the second largest trading partner with the U.S. as it has overtaken Japan as the largest Asian trading partner. 33% of all air freight movement between the U.S. and Asia is with China whereas Japan accounts for 24%. The growth of air freight has been so pronounced, many major Asian stations have either insufficient capacities or other inadequate infrastructure to support the continued growth of cargo activity.
Consequently, there has been a boom of infrastructure investment to support the continued reliance of air freight between the two regions. While the bulk of the growth has been on eastbound flights originating in Asia, westbound flights have also been expanding rapidly. From 1995-2005, Asian-North American routes have averaged a robust 7.8% annual growth while North AmericanAsian routes have only averaged a meager 2.6% annual growth. Their respective forecasts up to 2025 are 7.1% and 7.2%. This growth is projected in spite of exchange rate volatility which is widely believed to influence the volume of cargo shipped. There is no predominant type of cargo that is flown on eastbound flights. Boeing lists 59% of all cargo on these lanes as “other” while 11% are documents and small packages, 7% being electrical machinery, and 7% chemical materials. Westbound flights are also quite diverse. Office machinery and computes account for 16% of westbound cargo, 14% apparel, 14% telecommunications equipment, and 11% electrical machinery. The trade imbalance, however, has continued as eastbound flights carry substantially more volume than westbound flights. This has created problems for all players along the air freight chain, particularly freight forwarders. Other coordination disturbances have resulted which will be discussed in greater detail throughout this document. Commerce between Europe and Asia also has experienced substantial growth. From 1995-2005, westbound traffic has increased 12% annually while eastbound has experienced 4%. As is expected with North America, it is expected that growth over the next 20 years will average to be around 7% per year. North America – Latin America About 70% of all freight between North America and Latin America is with South America, mostly from Columbia and Brazil. Growth between these two regions has averaged 4.5% annually on northbound lanes and 0.3% of southbound. From 2005-2025, Boeing anticipates an average growth rate of 6.0% on flows into North America and 5.4% on flows out of North America. Most northbound flights carry consumer goods such as fish (23% of all northbound cargo), flowers (21%), and fruits and vegetables (11%). Southbound traffic contains mostly electronics goods lead by automated data processing machines (11%), manufactured goods and special machinery (10%), electrical machinery (7%), and industrial machinery and equipment (6%). Europe – North America Eastbound traffic has experienced an average annual growth rate of 2.5% since 1995, compared to westbound traffic which has been growing at 3.8% annually. Both routes are anticipated to accelerate with the expectation that eastbound and westbound traffic to grow at 5.1% and 5.6%, respectively. Most of the trade activity with Europe has been with Germany (21.4%), United Kingdom (19.6%), France (11.5%), Italy (9.3%), and The Netherlands (8.3%). On both routes, most cargo is considered to be documents and small packages, machinery, and specialized equipment.
North America – Middle East Despite the economic and instability in the Middle East, the region has been growing rapidly in terms of output. The United Arab Emirates, Jordan, Qatar, and Kuwait have averaged annual growth rates in excess of 6% from 2003-2005. This can be attributed mostly to the continued expansion of oil exports. Not surprisingly, air cargo activity has also been growing in the region and plays a role of increasing importance. North American cargo activity with the Middle East accounts for only 10% of all activity, but has been growing at 6.9% per year from 1995-2005 (and 7.8% from 2001-2005). Annual growth between the two regions from 2005-2025 is expected to be 7.6%. Eastbound lanes into the Middle East mostly consist of small packages, machinery, scientific equipment, and transportation equipment. Shipments westbound into North America mostly consist of apparel, fruits and vegetables, and pharmaceuticals. Figure 5 summarizes the average annual growth rates from 1995-2005 and the projections from 2005-2025 according to Boeing among major cargo lanes.
Cargo Volume: Historial and Future Expected Growth
1995-2005 2005-2025 (proj)
S.Am -> N.Am
N.Am - S.Am -> > S.Am Europe
Eur -> S.Am
N.Am > Eur
Eur -> N.Am
Asia -> N.Am
N.Am > Asia
Eur -> Asia
Asia -> Eur
D. Air Cargo Tomorrow
The fact that air cargo has grown in the presence of global economic volatility is indicative of the robust state of the industry. That said, there are a number of considerable challenges that the industry faces that provide a number of research opportunities. Among these:
1. Capacities – Many stations are reporting insufficient capacities to accompany the growth in the volume of air cargo. An example is at LAX, where the airport is having to locate storage facilities off-site to temporarily store cargo. 2. Security – Concerns regarding the security of air cargo and that of major cargo ports have been heightened post 9-11. However, inspecting 100% of air cargo is unambiguously an infeasible solution. Designing a security system allows for maximal inspection and implementing the best technologies to secure cargo while minimizing the distortion it induces on the efficiency is an open problem of great importance. 3. Fuel prices – According to the Bureau of Labor Statistics, the price of jet fuel has increased by 250% from 2000-2005 alone. Many believe the increased price level is structural, possibly experiencing similar increases moving forward. Figure 6 shows average annual price changes in jet fuel from 19752005. 4. Technology – The industry is becoming increasingly competitive and firms who adapt to the latest technologies are more likely to succeed – all else equal – than firms implementing antiquated technologies. Also, firms are increasingly relying on collaboration, so synchronizing technologies is imperative so firms are able to work together. As capacities become more limited, more freight is shipped, and competition becomes more intense, the demand to operate at the technological frontier is important to all firms. The ability to advance the frontier outward is even more important and always remains a challenge.
Change in the Price of Jet Fuel 1975-2005
% change from previous year
How the Process Works
Here we seek to understand the simple question of how a given piece of freight moves between its origin and destination. The general overview of the process can be summarized in the following figure:
Freight Forwarder Ground Handling at Departure Airport
• • • Pure freighter Integrator Combination Carrier no
Customs (if int’l)
Ground Handling at Arriving Airport
Freight Forwarder Consignee
A. The players involved
The following agents along the chain of air cargo are all crucial to the efficacy of the air cargo process. Let us formally define them. Shipper (or Consignor) – the one who requests service in transporting the cargo (source node of the supply chain) Freight forwarder (or forwarder) – analogous to a travel agent with passengers; the forwarder typically arranges for the transportation of cargo from the shipper’s warehouse, delivers (or has a contractor deliver) it to the departing airport, prepares the necessary paperwork, picks up at the arriving airport, and delivers (or contracts for the delivery) to the consignee. Carrier – The firm who provides the air delivery of cargo from the origin airport to the destination airport. There are primarily two classes of carriers: cargo carriers that primarily carry cargo and freight (e.g. FedEx, UPS, DHL, Kitty Hawk, Cargojet Airways), and combination carriers that carry both passengers and cargo that is stored in the bellies of aircraft (e.g. Korean Air, Lufthansa, Delta). About 25% of all cargo is carried on commercial passenger planes with the remainder on freighter aircraft. Ground Handler – An agent at an airport that physically handles the freight; this usually refers to whenever freight is loaded, unloaded, transferred, stored, retrieved, broken down, or consolidated. Consignee – The receiving party that the goods are sent to (sink node of the supply chain)
B. Process Flow: An Overview
Process flow of air cargo is subject to variability across stations. However, most stations follow the following overview. We will look at a case study in the subsequent section with a more precise analysis in Singapore. From the warehouse to the airport While shippers may deliver their cargo directly to the airport, the majority of cargo is handled through a freight forwarder. In most instances, the forwarder will pick the freight up from the shipper or the shipper’s warehouse. Unless the freight is rushed, it is generally brought to the forwarder’s warehouse and consolidated with freight to the same (initial) destination. Forwarders generally have their warehouses close to airports as freight is often transferred. The forwarder must also possess the air waybill which is a contract between shipper and carrier for the carriage of freight over the carrier’s routes.
Booking Days before departure, the forwarder will book freight on a given flight with a carrier. Most large forwarders receive a pre-allocated space on routes on a daily basis, but still call the carrier to specify the dimension and weight of the freight on a given flight for planning purposes. Forwarders may or may not have contracts with certain carriers. The booking process is becoming increasingly automated as it allows for more efficient operations and greater transparency. Generally the forwarder first reserves space, then confirms and picks up the freight only after receiving confirmation from the carrier. Consolidation Once in the hand of the forwarder, the cargo is likely to be consolidated with other shipments. The freight is likely to sit in a warehouse of the forwarder and await crossdocking, depending upon the priority of the goods. If the goods are not time-sensitive, they generally are stored one to three days in the forwarder’s warehouse to be consolidated before being delivered to the airport. Upon arrival at departing airport At the airport, the forwarder will deliver the freight to the carrier in which the freight goes through security. This is a significant component of the air cargo process and will be discussed in greater detail below. The air waybill is delivered, in which the forwarder possesses substantial penalty if there are discrepancies. The cargo is usually temporarily stored in a warehouse facility at the airport as it waits to be loaded. Usually, cargo is randomly assigned to bins which induce fast storage but slow retrieval (designing efficient storage mechanisms are of considerable interest). The cargo is then retrieved prior to arrival and loaded via a unit loading device (or ULD) and secured. The aircraft departs when all cargo is secure. At the arriving airport Cargo is flown either to the destination airport or an intermediate airport where it likely gets re-consolidated, re-containerized, and placed on another leg. Most likely, the cargo is temporarily stored at a warehouse where it is either transferred to a connecting flight or awaits to be picked up from the forwarder and delivered to the consignee. If re-consolidation occurs, there are a number of issues to classifying each piece of cargo. It can be consolidated by: size and weight, special requirements (e.g. refrigerated goods together, dry ice cannot move with live animals, etc.), service level (by priority), or destination. Different operators may consolidate differently. If a given leg is international, then the cargo is subject to customs clearance which is done at the airport. Delivery From the airport, a forwarder typically will deliver the cargo in the reversed order from where the forwarder delivered the cargo to the airport.
The air cargo process is driven by the freight forwarder, the carrier, and the process flow. Here we study each of these principal components separately.
C. Process Flow: A Case Study at Singapore Changi Airport
Due to the aforementioned variability in the process flow across stations, let us examine the precise flow at Singapore Changi International Airport (SIN). SIN is one of the busiest cargo airports by volume globally, and is the hub for Singapore Airlines, one of the leaders in freight transportation. Singapore Airport Terminal Services, Ltd (or SATS). is the official terminal operator in Singapore, and one of the most reputable. SATS handles over 75% of all of the air cargo flow in Singapore. Opened in 2001, SATS Airfreight Terminal 6 is one of the most sophisticated global airfreight facilities which required a $270 million investment. Please note that information of the process and information flow at SATS is from the Bazaraa, Hurley, et al. paper (2000). Shipper Booking with Forwarder The first obvious step in the process is for the shipper to contact the forwarder (unless the carrier that will transport the cargo is an integrator, these will be different entities). Orders are placed either through faxes or phones. The forwarder then will reserve space with the carrier, and only after the carrier confirms space with the forwarder are arrangements made to pick up the freight. If the forwarder has pre-allocated space from a contract with a carrier, the confirmation is almost immediate. Otherwise, the confirmation will take substantially longer. For international shipments, the forwarder needs to apply through Tradenet for an Export Cargo Clearance Permit, as well as other specialized permits that depend on the cargo being shipped. In most cases, the permit is approved within 30 minutes, but the forwarder is generally allowed the obtain the permit within three days after the export. Once booked, the forwarder will send an itinerary to the shipper, and trucking is scheduled for pick up. The cargo is then picked up, together with the necessary paperwork. Generally the forwarder will make multiple pick ups before returning to the forwarder’s warehouse where cargo is labeled and processed to the export department. Once this is processed, it is officially measured and weighted for the air waybill to be completed. The freight is stored, a pick list is generated, and the cargo is loaded to be brought to the airport. Check-in/Registration Upon arrival to the airport, the forwarder clocks in the control form and proceeds to the acceptance counter. There is an acceptance officer on duty who first process all the documents upon accepting the cargo. Weighing is then done and cross-checked with the AWB.
As is with the case of most facilities, lose cargo induces inefficiencies. Loose cargo needs to be consolidated or built up at the facility which requires labor. Because there are approximately 250 forwarders in Singapore, none with excessive market power, lose cargo tends constitute a greater share at SATS than at most other facilities. Ground Handling to Departure Meanwhile planners monitor activity for each future departing flight and particularly monitoring arriving cargo that needs to be transferred within a short time frame (generally two to six hours). Planners have a list of the containers and a loading plan, as well as a loading priority list that is to be loaded first. In the presence of overbooked flight legs, planners will consult the airline for route management issues. • For loose cargo For cargo accepted within four hours of departure is moved to a build-up area and placed on a pallet or in a container. Cargo accepted more than four hours prior to departure is placed in a storage bin that will be placed into the Automated Storage/Retrieval System (ASRS) once it is full with other freight. The storage location is logged and later retrieved. Twelve hours prior to departure, the export retrieval team will sort cargo into bins according to flight number. About four hours before departure, cargo is built up in a location where all cargo is built up. A label or tag is then placed on the consolidated cargo and is transferred vis-à-vis an elevating transfer vehicle (ETV) to the pre-load area near the aircraft. This is done as early as two hours before the departure time. The master loader typically begins loading cargo from this area an hour prior to departure. As cargo is moved toward the pre-loading area, the flight manifest is updated until the completion of all built-up cargo for the flight leg. This manifest is a master list of all cargo on the given leg and contains the AWB numbers, weight, origin and destination, and a description of a good. • For palletized/containerized cargo The process flow for cargo not required to be built up is similar to that of loose cargo. The only difference is that pallets or containers do not require further consolidation, and are not brought to the general consolidating area. If built-up freight is checked prior to four hours, a label is immediately placed on the container and are brought to the preloading area. Otherwise, they are stored in the automated warehouse system and will be retrieved before departure. • For specialized cargo For cargo requiring special needs, there are specialized facilities located throughout the airport that such cargo is brought to. Arrival to Ground Handling Generally, the arrival process is just a reversal of the departure process. Cargo is brought to the terminal warehouse office and sorting begins. AWBs and other documentation are sorted by accounts. Then they are placed in pre-specified pigeonholes according to the
forwarder contracted to pick the cargo up if the arrival airport is the destination. The consignee may be notified of the arrival if requested. While the cargo is being unloaded, documents are processed which are surprisingly time consuming. Cargo from passenger aircraft takes approximately two hours to complete and three hours on a freighter aircraft. Off-loading generally takes place between 30 and 45 minutes after arrival and all cargo is then verified. For cargo requiring transferal, it is sent to an export station and it goes through essentially the same process as at the arrival airport without the registration. Special cargo is brought to the specialized facility. Ground Handling to Forwarder In most cases, the forwarder will receive pre-notification of arrival. Cargo is generally ready to be picked up two hours for passenger aircraft and three hours for freighters. The forwarder collects the air waybills and related documents, and a delivery order is generated. Through TradeNet, import cargo clearance permits and other various permits (depending on the contents of the cargo) are applied for, and issued within 30 minutes. After clearance, the warehouse department retrieves the cargo the transportation department begins planning truck delivery routing. The cargo is then given to the forwarder, in which the forwarder checks the condition of the cargo and delivers it to its warehouse. It is then either broken down or sent directly to the consignee.
III. The Freight Forwarder
The freight forwarder plays an integral role in the air cargo process. While they essentially are responsible for coordinating the movement of freight between the shipper and consignee, the mechanics by which they operate and the environment under which decisions are made is complex. We seek to describe this here.
A. Classes of Forwarders
Before proceeding, it is important for us to identify different classes by which forwarders are characterized. Forwarders are coarsely grouped into two types of firms: integrated and disintegrated. Integrated forwarders play the role of both forwarder and carrier; their firms operate at each stage of the air freight process. That is, they pick up, transport, consolidate, book, carry, and deliver freight. Disintegrated firms, however, do not carry freight. At a minimum, disintegrated firms only coordinate the movement of freight; they contract the delivery process. However most forwarders provide the trucks (either directly or through a contractor) and deliver to the departure airport and pick up at the destination airport.
B. Market Facts
In 2004, the global freight forwarding industry amounted to $27.63 billion. Not surprisingly, the freight forwarding industry has grown with the air freight industry. An interesting observation is to note that there has been more market concentration in the industry. Deutsche Post alone, the German postal service, alone has made over 100 acquisitions since 2003 alone including all of DHL, 25% of Lufthansa Cargo, Air Express International, Airborne Inc., and SmartMail. Among the major acquisitions recently, in 2005, Deutsche Post acquired Exel for $6.6 billion. The next year, Deutsche Bahn acquired BAX Global for $1.2 billion. The most common explanation for this increase in concentration is the fact that the high growth in cargo drew more forwarders into the industry at first. Many of the firms that got bought out were nitsche providers. In general, coordination among firms abates double marginalization and such a mechanism cannot be sustainable in an industry as competitive as the freight forwarding industry when margins are low to begin with. Federal Express (FedEx), United Parcel Service (UPS), and DHL account for the three largest forwarders, often collectively referred to as the “Big Three”. These are all integrated forwarders, but are somewhat different from a traditional freight forwarder in the sense that the majority of their volume accounts for small packages which do not account as “freight”. Each of these firms has separate divisions distinguishing documents and small packages apart from pure freight.
Table 1 shows some data of these three giant firms, as well as the size of their air and ground fleets. Table 1 Summary of FedEx, DHL, and UPS (2006 data) % of total revenue FedEx DHL Air freight 10.96% 6.56% Forwarding services/logistics 11.75% 20.33% (includes small packages) Aircraft Freight Vehicles 671 42,000 420 76,200
UPS 1.88% 11.13% 268 91,700
Besides the so-called “Big Three”, other dominant freight forwarders include Nippon Express, Kuhne & Nagel, BAX Global, Schenker, Kintetsu, and Panalpina. These firms, along with FedEx, DHL, and UPS accounted for about 70% of the total global air freight forwarding in 20043.
C. Case Study: BAX Global
To illustrate how a typical freight forwarder operates, let us study one of the major global freight forwarders, BAX Global. About BAX Global • • • • • Founded in 1972 as Burlington Northern Air Freight World Headquarters: Irvine, California $.29 billion in revenue in FY 2005 (77% is from international transport or logistics service) Currently operates about 500 facilities in over 130 countries In January 2006, Deutsche Bahn acquired BAX Global; the value of the company increased from $3 billion to $12 billion.
Services Provided In addition to air forwarding, BAX Global provides the following services: • Ocean freight forwarding • Trucking services • Consulting • “Value Added” service • Warehousing
Source: 2005 Deutsche Post Annual Report
Given the recent increases in the price of air freight transportation, ocean freight has become relatively less expensive for the forwarder (and hence, the shipper). Consequently, many forwarders are focusing on growing their ocean forwarding services as it has traditionally not been as well developed as air freight forwarding. The fact that firms face increasing demand for ocean forwarding helps explain the increased market concentration. As mentioned, in January 2006 Deutsche Bahn (the German logistics provider who also owns DHL) acquired BAX Global from The Brinks Company (the former parent company of BAX Global) partly to harmonize their global logistics profile. More specifically, Deutsche Bahn acquired Schenker Logistics who has been a dominant ocean forwarder. The further acquisition of BAX Global is likely a result of this change in the composition of forwarding services. Logistics at BAX Global Domestic freight is either shipped by truck or plane, depending on the priority of service requested by the shipper. The truck fleet is operated by a number of trucking firms who are contracted (e.g. Schneider, ATi, USPS, UPS, among others), some of whom are required to display the BAX logo on their trucks. There is a network consisting of six “Gateways” and six consolidation hubs. With respect to air freight, there is a single hub in Toledo, Ohio. The fleet consists of 21 DC-8s (which, as of December 2006, constitutes as the largest operator of DC-8s), with about three aircraft being in maintenance at any given time. Freight also may be transported by commercial air carriers who BAX contracts with. While BAX owns its own fleet, they fly only domestically. Thus, they are most aptly characterized as an integrated fleet with respect to domestic cargo and a disintegrated fleet with respect to international freight. The fact that there is a considerable imbalance between outgoing and inbound freight poses challenges between the coordination of them and their contract carriers (see section 1 for a description of this imbalance; this is particularly pronounced on Asian-North American routes). Because of this imbalance, freight forwarders exhibit bargaining power with carriers on North America to Asia lanes, whereas the carrier will exhibit such power on Asia to North America lanes. To hedge this imbalance, BAX Global guarantees its shippers that for every container that is outbound to Asia, they will accommodate four containers inbound from Asia. Supply Chain Coordination Collaboration between freight forwarders and other global supply chain participants is critical to the success of firms. Prior to 2003, BAX operated mostly independent from their competitors in their trucking network. At the beginning of that year, on most lanes they made their trucks available to other forwarders to enable them to (a) expand their existing network and (b) improve upon load factors to their trucking fleet.
Coordination between participants along a supply chain becomes increasingly important given that margins remain to be low for the forwarders. Contract design in the freight forwarding industry (as well as that for the air cargo industry) remains a topic of interest that is not well developed. This topic is considered one of the more open questions.
IV. The Carrier
Clearly in order to study airfreight, one must study the carriers. We begin by listing the major carriers, and characterize the major firms. Then we study how cargo operations differ among the major firms. We also study airports and how capacity constraints and major facilities may influence the future of the air carrier.
A. Types of Carriers
Air carriers can principally be decomposed into three classes: freight carriers, combination carriers, and integrators. Freight carriers are devoted solely to freight, while the combination carriers include airlines or subsidiaries of large air carriers who carry passengers while storing freight in the belly of aircraft. Integrators oversee the entire process and act as the forward and the carrier. Examples of freight carriers include Cargojet, Kitty Hawk, Gemini Air Cargo, Polar Air Cargo, among many others. Some airlines have subsidiaries that act as a freight carrier, but are likely to be considered a combination carrier, like Lufthansa Cargo, a subsidiary of Lufthansa Airlines that carries only cargo, while Lufthansa Airlines is clearly a passenger airplane. Combination carriers are generally passenger aircraft that additionally carry cargo aside. Cargo used to be a byproduct of passenger airlines, but has grown in the importance of most firms’ operations. However the size and importance of cargo operations greatly differs across firms (see below). So Lufthansa also is likely to carry freight in any given flight in an A-320 which makes Lufthansa Airlines a combination carrier. However their subsidiary Lufthansa Cargo is a freight carrier. In the U.S., Northwest Airlines (as of 2006) was the only U.S. airline that operated freighters. The primary integrators are the big three: FedEx, DHL, and UPS. A number of estimates show that with regards to all air freight flown by combination carriers, roughly 75% is still flown on freighter aircraft while the remaining 25% is integrated with passenger planes. It is interesting to note that in their annual reports, the big three have indicated a movement away from small packages and documents to larger freight. This could have a profound impact on the industry as these giants exhaust a greater share of large cargo transported. Another noteworthy trend is the development of so-called niche providers. While there is considerable market concentration in the airfreight industry, the number of firms has still been increasing. One explanation for this observation is the increase in the number of carriers who specialize in specific types or classes of freight transportation. Many of these are typically smaller firms who are contracted to fly for larger carriers. So while the 22
number of carriers has been increasing, the aforementioned major carriers have been transporting a greater share of all freight transported. Examples of niche providers include those firms that specialize in the transportation of: produce, garments, live animals, wine, forestry products, IT equipment, medical supplies, frozen/refrigerated goods, among other specialized freight.
B. Major Carriers
Table 2 lists the largest carriers of air freight overall in 2004. Please note that for combination carriers, the FTK data represents the sum of freight flown in passenger aircraft and freight flown by their subsidiaries. Table 3 shows some data on load factors, and Table 4 shows a sample of larger passenger carriers and the share of profit from cargo. Table 2 Freight-Tonne Kilometers (FTKs) Flown, 20044 rank Carrier Millions FTKs 1 FedEx 15.58 2 Korean Air 8.26 3 Lufthansa Cargo 8.04 4 UPS 7.35 5 Singapore Airlines 7.14 6 Cathay Pacific 5.88 7 China Airlines 5.64 8 Eva Airways 5.48 9 Air France 5.39 10 Japan Airlines 4.92 Table 3 Sample of Load Factors at Major Carriers5 Carrier Load factor Korean Air 76% China Airlines 73% Lufthansa Airlines 65% Singapore Airlines 64%
Source: International Air Transport Association Source: each carrier’s annual report (2005)
Table 4 Sample of Cargo Share of Revenue at Major Combination Carriers6 Carrier Cargo Revenue, as % of Total Revenue China Airlines 43.54% Korean Air 32.28% Singapore Airlines 31.49% Cathay Pacific 28.50% Continental Airlines 8.66% United Airlines 4.30% Table 4 verifies that the share of cargo at major combination carriers is subject to considerable variation. A natural question to ask is what a break-even load factor must be. While data is not readily available, according to their annual reports, China Airlines and Singapore Airlines reported break-even load factors of 72% and 63%, respectively.
It is naturally of interest to study airports as they have a great impact on the efficacy of the air cargo process. A natural place to place to begin is identifying the major airports utilized in airfreight. Table 5 exhibits the 20 busiest stations of air cargo handled by volume in 2005. A great concern is that whether airports, particularly the ones listed in Table 5, have sufficient capacity to accommodate the expected growth in air freight even in the short term. An example is at LAX which serves as an important gateway to and from Asia. The facility is so capacity constrained that other nearby facilities are being utilized more. For example, John Wayne Airport, Ontario, and Long Beach are airports that are being increasingly utilized as a result. According to the Ontario airport authorities “the owner and manager of both (LAX and ONT) is counting on ONT to assume tremendous cargo growth as LAX confronts critical capacity limitations7. This problem reiterates the importance of planning considering that the volume is not expected to be constant, but it is anticipated to grow at 6.1% per year from 2005-2025 (according to Boeing). One area that has grown that has been able to increase throughput at some facilities has been the implementation of automated retrieval systems. Further work in picking, storing, and cross-docking is important to increase efficiency and therefore the ability for facilities to accommodate the future anticipated growth. This is an important challenge for airports, particularly consolidation hubs, and will be discussed in greater detail below.
Source: each carrier’s annual report (2005) Source: www.ci.ontario.ca.us/index.cfm/2559/3817
C.1 HACTL One of the most important facilities in the air cargo world is Hong Kong. It is also one of the most efficient. Hong Kong Air Cargo Terminals Ltd, or HACTL, is largely considered a model for how a cargo facility should operate. It is unambiguously considered the most technologically advanced facility. Because it is such a major player in global air cargo, let us look at some basic facts before looking at why it is considered a model facility. • The SuperTerminal 1 at HACTL is the world’s largest stand-alone air cargo handling facility. • Opened in 1976, SuperTerminal 1 recently opened after an investment of more than $1 billion. • It has an annual capacity for 2.6 million tonnes of capacity. • The land area for SuperTermianal 1 alone is an astounding 170,000 square meters. The reason why HACTL is largely considered the model facility can be most succinctly characterized in one word: automation. Let’s look at some of these components8. • The container storage system, for example, is fully automated with more than 3,500 storage positions. There are 12 computer-controlled stacker cranes for moving cargo efficiently within the network. • The box storage system (BSS) contains more than 10,000 bulk storage positions for individual consignments. Goods move in and out of BSS with no human intervention. • Even cargo unit and pallet handling are also highly automated. Units and pallets go directly from the tarmac into the storage system which has more than 3.500 storage positions. • The top floor of the six-story facility is dedicated to breaking down and building up freight according to a computer-generated plan in one of over 300 individual workstations. This is done in a tightly-restricted highly secure area. Containers move throughout the facility by a highly sophisticated level of conveyor loops and computer-controlled cranes. • The security system is also considered the most advanced. It has consistently had the lowest rate of pilfering among major ports. While many successful cargo facilities have implemented information technology that lies near the frontier of available technologies, HACTL is still considered a model for major hub facilities, and it ought to be emphasized when studying the air cargo industry. C.2 Busiest Airports by Volume Table 5 shows the 20 busiest cargo airports globally (2005 data). 8 of the 20 are U.S. facilities, 7 are Asian facilities (3 of which are in China), 4 are in Europe, and 1 in the Middle East (Dubai).
Source: HACTL’s website at www.hactl.com
Figure 7: Container storage systems at HACTL
Figure 8: Bulk cargo distribution system at HACTL (sushi-style loop conveyor system
rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Table 5 Busiest Cargo Airports (2005)9 Station Metric tonnes handled Memphis International (MEM) 3,598,500 Memphis, Tennessee Hong Kong International (HKG) 3,433,349 Hong Kong, China Ted Stevens Anchorage International (ANC) 2,553,937 Anchorage, Alaska Narita International (NRT) 2,291,073 Narita, Japan Incheon International (ICN) 2,150,140 Seoul, South Korea Charles DeGaulle International (CDG) 2,010,361 Paris, France Frankfurt International (FRA) 1,962,927 Frankfurt, Germany Los Angeles International (LAX) 1,938,430 Los Angeles, California Pudong International (PVG) 1,856,655 Shanghai, China Singapore Changi (SIN) 1,854,610 Changi, Singapore Louisville International (SDF) 1,815,155 Louisville, Kentucky Miami International (MIA) 1,754,633 Miami, Florida Taiwan Taoyuan International (TPE) 1,705,318 Taoyuan, Taiwan John F. Kennedy International (JFK) 1,660,717 New York, New York O’Hare International (ORD) 1,546,153 Chicago, Illinois Amsterdam Schipol International (AMS) 1,495,919 Amsterdam, The Netherlands London Heathrow International (LHR) 1,389,589 London, England Dubai International (DXB) 1,314,906 Dubai, United Arab Emirates Bangkok International (BKK) 1,140,836 Bangkok, Thailand Indianapolis International (IND) 985,457 Indianapolis, Indiana
Source: International Air Transportation Association
Here we examine aircraft used my major freighters. For combination carriers, freight is stored in the belly of the aircraft so aircraft utilized on these routes are of little interest. Therefore we restrict our attention towards freighter aircraft. D.1. Overview Many smaller freighters operate retired aircraft. For example, in 2006 BAX Global was the world’s largest operator flying 21 DC-8 aircraft. Kitty Hawk Aircargo is one of the larger operators of the antiquated Boeing 727 fleet while operating 28 such aircraft (26 of which are pure freighters; the other two are Boeing 727-200; they also operate 7 Boeing 737-freighters). Yet most of the firms that operate retired commercial aircraft or old freighter aircraft constitute a relatively small share relative to their larger counterparts. Let us examine some of the aircraft statistics of freighter aircraft operated by these large entities. D.2. Freighter Data Overall, Boeing leads the freighter industry overall. According to Boeing, their aircraft provide “over 90% of the of the total worldwide dedicated freighter capacity”10. Boeing 747-F Boeing 747 freighters exhaust two-thirds of the world’s wide-body freighter fleet11. Further, it is estimated that half of all global airfreight is carried in a B747-F. Most the larger carriers operate a B747-F in their fleet, particularly for trans-continental routes. Payload capacity is in excess of 124 tons (113,000 kg) and the standard range is 4,450 nautical miles (extended range versions can travel an additional 525 nautical miles). Maximum takeoff weight is 910,000 pounds.
Figure 9 (source: cargo.koreanair.com)
Boeing 777-F Launched in May 2005 the Boeing 777-F was introduced specifically for long-hauls with a maximum range of 4,885 nautical miles (9,047 km). The twin-engine aircraft was marketed as inducing the lowest unit operating cost. Payload is slightly less than the 747F at 114.5 tons (103.9 metric tones). Boeing insists that this particular aircraft will induce the lowest unit operating costs as a result of the increased range. Boeing 767-F While older than the B-777 F, the 767-F remains the world’s second most popular freighter aircraft which is primarily used a medium-size aircraft. It is marketed as being robust and fuel efficient. Over 40 airlines own a B-767 and is crew-compatible with the B-757 freighter. There is a two-person flight deck and its twin-high-bypass ratio is said to provide efficient fuel economy. Unlike the B-747 and B-777, it is not primarily used as a long haul freighter with a range of 3,270 nautical miles (6,056 km). Payload capacity is approximately 60.5 tons (54.88 tonnes). The aircraft is good for bulk loading as it will hold virtually all types of containers. The only real competitor to the Boeing family is the Airbus freighters which are a distant second in terms of share of global air freight. However, many industry analysts believe that the introduction of the Airbus 380 could increase Airbus’ share of the freighter market. However, in early 2007 Airbus announced that it would delay the production of it’s 380F version. This could have a more global impact. UPS was scheduled to purchase several A-380 F models, but has canceled several of these in March 2007 as Airbus has reported that it would be shifting several of their production staff to the development of the A-380 passenger model. Despite Airbus’ relatively small market share in the global airfreight market, there are still a number of aircraft that transport a substantial volume of freight, most notably from FedEx, DHL, and Lufthansa Cargo. Airbus 300-F (A300-600 F) The A300-F is designed primarily on shorter routes with a range of 2,650 nautical miles (4,850 km) and bulk holding area of 298 cubic feet. The double side-by-side loading area allows for a capacity of 21 pallets, while the single row loading design allows for 15 pallets.
Figure 10 (source: cargo.koreanair.com)
Oversized Freight: Beluga (Airbus) Designed to carry oversized cargo like airplanes, helicopters, and other aircraft, as well as other abnormal sized cargo, Airbus designed the A300-600ST as a wide-bodied version of the A300-600F. With a length of 184 feet, span of 147 feet, and empty weight of 86 tons, the A300-600ST, or “Beluga” as it is commonly known (see Figure 11 for an evident explanation as to why it is frequently called this) is one of the largest – and most innovative – of aircraft designs. The Beluga has a cargo volume of 848 ft3 (or 1,365 m3), cargo capacity of 47 tons, and requires a cockpit crew of four to operate. The aircraft replaced previous so-called “Super Transporters” (hence the ST in the aircraft model) including the “Super Guppies” that began in operation in the early 1970’s.
Figure 11: Airbus A300-600ST or “Beluga”
D.3 The Air Freighter of Tomorrow Noise restrictions are also sensitive to the industry considering the bulk of transcontinental activity is conducted around midnight where there are more restrictions. 30
Another issue that is heavily pursued is sustainability and implementing policies that minimize unnecessary emissions. This is currently of interest to many that study the industry. Among the more notable studies that has been active in the community is a socalled “continuous descent” approach whereby aircraft climb to a given ascendancy point quicker and descend in a more gradual, continuous, manner instead of a more rapid descent. The advantages to a continuous descent approach include: • • • Increased cost effectiveness Emission reduction Noise reduction
There was a trial at Louisville International Airport (SDF) which serves as the main hub for UPS in 2004. Although the approach is subject to variation in pilot ability, wind speed, and aircraft weight, the conclusion validated the preceding three advantages. CO emissions below 3,000 feet were reduced by about 15% (depending on the aircraft); HC below were reduced by about 18%; and NO2 emissions below 3,000 feet were reduced by 34%. Further, the fuel burned to fly the last 180 nautical miles to the runway was reduced by 364 and 118 pounds per flight on a B-767F and B-757, respectively. The time it took to fly the last 180 nautical miles also declined by about two minutes12. Studying continuous descent approaches are one of many mechanisms that are being studied to improve both cost and environmental considerations. It has gained traction among those who study the industry, and is possible to become increasingly a focus point moving forward. D.4 Containers As the shape and size of freight is subject to variation, there are also many different containers to accommodate different classes of freight. The most common types of containers will be owned by all airlines. Most bulk freight comes in either a closed container or is shipped on an open pallet. Containers are also tailored for specialized freight. For instance, there are refrigerated containers, horse stalls, and other specialized containers. The most commonly used unit load devices (or Olds) are the following: LD3 The LD3 container is the most common general closed containers (see Figure 8). The volume of the LD3 is 153 ft3 (or 4.3 m3). The larger freighter aircrafts generally have a capacity for at least 30 such LD3 containers. A Boeing 777 has capacity for 30 LD3s; An A340-300 has capacity for 32 LD3s, and an A380-800 has a capacity for 38 LD3s. There is also a refrigerated version of the LD3 for produce and freight requiring temperature control.
note: noise was reported to be greatly reduced but was not quantified due to unexpected wind changes
Figure 12: LD3 Container
Pallets Pallets are just rectangular bases that support open faced cargo that is secured. Pallet sizes differ; the most commonly used pallets are of (approximate) dimension 7’x10’x8’ that have a maximum tear weight of 275 pounds (125 kg). The other is of dimension 7’x10’x5.5’ with the same max tear weight (see Figure 9).
Figure 13: Pallet
E. Freight Carrier Economics
The freight carrier industry is becoming increasingly competitive, and therefore a highly complex environment. Here we study some aggregate industry-wide facts of the industry, examine market concentration, how some planning decisions are made, and how carriers coordinate. We conclude by a case study of air cargo revenue management implemented at KLM. E.1 Coordination Given the number of players and the multiple stages of the air cargo supply chain, coordination becomes an important means of a carrier’s strategy. Also, we will see that the industry has become increasingly concentrated. We begin by first studying where carriers coordinate with other firms. I focus primarily on the coordination between the carrier and the forwarder as these two players do most of the handling on a given shipment. Carrier-Forwarder Coordination 32
Recall the case study of BAX Global. While BAX has a fleet that flies within North America only, they contract with certain carriers on transcontinental routes. Such collaborations are favorable as they abate the asymmetries between certain routes. We have seen that on trans-Pacific routes, forwarders exhibit market power on flights to Asia while the carriers have market power on flights that originate in Asia. Another critical point in carrier-forwarder coordination is risk sharing. Technologies are making it increasingly possible for the coordination to improve. Tracking technologies are centered at RFID that has allowed customers, carriers, and forwarders to coordinate together. Considerable resources have been, and continue to be allocated toward improving this further so as to make cargo technology as transparent as possible. One considerable barrier the industry faces is integrating the technologies of different firms. Since Deutsche Bahn acquired them, BAX Global expressed a great challenge of integrating their IT systems with that of their partners and has expressed how the frictions between incompatible systems have not allowed full collaboration to take place. Forwarder-Forwarder Coordination While shippers use forwarders as an intermediary in the delivery of freight, the forwarder contracted by the shipper may not be the only firm that is used in ground transportation. Shippers often collaborate with other shippers for two reasons. The first is asymmetry in forwarder’s network. If at a given receiving destination the carrier contracted by the shipper has little presence, the forwarder will generally have the service provided by a different firm who collaborates with the forwarder. For instance, BAX Global has a fleet of trucks with the BAX logo on, but are neither owned nor operated by BAX. They also contract with UPS, FedEx, USPS, among other forwarders. The second reason for establishing coordination with other forwarders is to improve load factors. Naturally spoilage (excess capacity in this context) occurs more frequently when a firm operates in isolation than when they are able to transport some cargo for other firms. An example comes from the trucking industry. Many of those familiar with the industry cite the asymmetry involving truck loads in and out of Florida; freight flowing into Florida outweighs that of flowing out of Florida, and given the geographic isolation from Florida, hedging this asymmetry becomes an important policy for many trucking firms. Carrier-Carrier Coordination Some carriers will coordinate with each other, motivated by the same principle by which forwarders coordinate, namely risk sharing. DHL, for example, coordinates with several other non-integrators to carry some cargo on specific routes. FedEx and UPS, however, carry almost all their freight on their own fleets. E.2 Pricing Another question that naturally arises is how pricing is determined. Unfortunately there is not a precise answer to this complex question. What we do know, however, is that pricing
is usually state-dependent as we expect any firm that uses revenue management to exhibit. In the case of the freight forwarder, pricing is based according to volume, priority, service, and other characteristics. As we would expect, consolidated shipments are charged separately in which the shippers’ freight is consolidated (generally at the shipper’s warehouse) and are charged a lesser price. Non-consolidated (LTL) shipments are considerably more expensive. With respect to the pricing of the carrier, we will study the complicated issue of revenue management below. Forwarders either book freight for individual flights, or have contracts with carriers in which the carrier reserves the given allocation on each flight with the forwarder. For flights in which the forwarder book separately, prices are subject to volatility. Of course the price depends upon the capacity (and forecasted capacity), dimensionality of the freight, and other factors. It may be of interest to learn about the aggregate behavior of price indices. Figure 14 shows the price indices for inbound versus outbound freight in the U.S. from 1992-2006. We note two salient observations: first, the price of airfreight is subject to volatility. Secondly, which should not be surprising considering the behavior of the volume of air freight, that the price of both inbound and outbound freight have risen considerably since 2002 (recall that 2001 was a recessionary year). Moreover, the price changes for inbound and outbound freight appear to have grown at the roughly the same pace.
Price Indices for International Freight 1992-2006
outbound flights inbound flights
80 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
In the past, pricing would be generally static – there would be a fixed charge for a given service that would rarely adjust. However technological advancements have made pricing more dynamic in the air freight industry. With market segmentation, pricing is an extremely complicated and highly-dimensional problem. This problem is of extreme importance and needs to be studied in greater detail. We now turn our attention to this problem of revenue management. E.3. Air Cargo Revenue Management In the passenger airline industry, revenue management became an extremely useful and important tool used by airlines in the 1990s. The idea is to charge the right price to the right person at the right time so as to maximize revenue. Passenger revenue management is highly dimensional as time of booking, type of passenger (e.g. business versus leisure passenger), origin-destination pair; among other factors is a function of the price of the ticket when issued. Despite this complexity, advancements in computational technology have made these previously unsolvable problems tractable and have become a vital component to any carrier. Revenue management is also used with cargo carriers, and the idea is nearly identical to that of passengers. As with passengers, cargo space is allotted in advance or sold for free-sale. Both strategies use an iterative schedule planning process: first a central planner schedules a draft that is later evaluated in the second phase where it is modified if needed. Both are driven by “demand driven dispatch” – the type of aircraft is subject to change as a function of demand. Despite these similarities there are also considerable differences, and the problem of revenue management of air cargo is even more complex regarding passengers. The following distinguish these complications: • Unlike passengers, cargo can easily be off-loaded subject to on-time delivery. Consequently, the number of possible ways to route a given shipments exceed that to route a passenger. • In passenger revenue management, each entity is well-defined: a passenger. In cargo revenue management, each item is more complex: length, width, and weight now define each entity. • Many bookings in air cargo are cancelled, re-booked, and cancelled again. Therefore, the booking process is subject to considerable volatility. • Demand is lumpy and forecasting future demand is thought of as being at least as complicated as passengers. Despite the difficulty in analyzing revenue management in air cargo, the potential benefit is far too ignoring the complexity in studying this. There has recently been some work done in the field that has suggested various policies so as to improve revenue management. Before looking at these, let us first decide the challenges in modeling. The first is how to abate the curse of dimensionality. Even with modern technology, solving such highly dimensional problems is infeasible and obvious assumptions and simplifications are inevitable. Another is whether the carrier is assumed to be a combination carrier or a pure freighter. Different booking policies and pricing strategies would be employed under each scenario. Overage costs are vital to cargo because rebooking is a more viable option in cargo as oppose to passenger (i.e. overbookings
would naturally be far greater in the cargo framework as a result). Spoilage costs are also critical to designing policies of a carrier. It has been shown (see below) that that the overage/spoilage costs are essential in governing revenue management policies of a carrier. Another complication is modeling show up rates. Due to the high fluctuations in booking and rebookings (mostly by the freight forwarder) the show up rate is volatile and also seen to be essential in appropriately modeling air cargo revenue management. The differences in revenue management between passenger traffic and air cargo is so vast that there have been papers published just to enumerate these. As with the case of treating passengers as entities, there are principle models that capture the decision process of booking cargo: Markov decision process (MDP), bid price accept/reject decision, and dynamic stochastic knapsack problem. The last of these is not as common in the passenger model as entities are well defined. There are clear advantages and disadvantages to each of these. Most of the papers listed below in a brief survey share the same characteristic in the sense that cargo is often offloaded, re-routed, and is subject to many changes given the low relative cost of overage relative to the high cost of spoilage (of course this is not the case in high-priority freight, but the routings of these are more exogenous). Even by making generous assumptions regarding ways to abate the dimensionality problem, the resulting problem is still hard to find. Therefore, the next question is how to apply a good heuristic to find a good, if not optimal, solution. Amaruchkul, Cooper, and Gupta (2005) simulate seven heuristics by looking at capacityto-demand ratios with respect to both weight and volume. Appropriate bounds are easily found and deviations of the outcomes with respect to each policy’s upper bounds are compared. The best heuristic depends centrally on the capacity-demand ratio. Not surprisingly, a higher penalty cost induces a more conservative booking policy in a simple accept/reject framework. Perhaps most notably, the authors show that airlines can boost revenue substantially by reducing uncertainty of their volume. This is somewhat surprising considering the aforementioned fact that overbooking is seen less costly in air cargo. However, the authors’ result suggests that airlines may overbook too much in which the tradeoff of higher volume of shipments is more than offset by the reduced reliability of the carrier. Therefore, it may be of a shipper’s interest to implement policies so as to reduce the volatility of volumes. For instance, they may wish to introduce a more standard shipping size.
Case Study: Revenue Management at KLM While air cargo has made some advances in the academic literature, implementation is a concern in which the recommended policies are not necessarily tangible. The way airlines use revenue management and the way the literature argues they should do not necessarily coincide. Fortunately this issue has been documented in a paper by Slager and Kapteijins (2004) in looking at a case study of revenue management at KLM.
Forecasting cargo is much more difficult than passengers because of the lumpiness of demand as well as the erratic behavior of the cargo booking process. Moreover, KLM is a combination carrier so balancing passenger and cargo bookings is a challenge and complicates revenue management. Air cargo revenue management at KLM is not as new as some may believe. In the mid1990’s KLM Cargo initiated a program called “System Profit Management”, or SPM. However, the program ultimately failed because it “had focused too much on IT and a scientific approach, instead of the numerous complexities of the day-to-day process.” (Slager and Kapteigns, pp. 82). In 2001, revenue management was brought back into what they coined as “margin management” (MM) which sought to implement the conceptual gains made from SPM. The goal of MM was to optimize the margin of cargo by (a) increasing revenues (b) decreasing the network costs for handling and (c) enhancing load factors in both volume and weight. To continue the implementation of MM, two organizations existed: MM Execution which was responsible for the day-to-day execution of MM process in each of the three areas and MM Process, and MM Process & Tool Development which was responsible for the development of systems in which each of the key players in their major markets could access. Margin management continues to prosper at KLM Cargo today. As with most carriers, cargo is sold one of two ways: 1. Contract – Here there is a mutual agreement between KLM Cargo and their customers (i.e. forwarders) in which the customer is guaranteed a certain capacity on a given flight leg. 2. Free Sale – Also called “Request/Reply” or R/R bookings – Here space is sold prior to departure with no guarantee on capacity. R/R can be sold at either: list prices that are published in advance, customer specific prices, or order specific prices. The revenue management of each of the two cargo products differs. Let us examine each. • MM on contracts Twice per year, at the start of the IATA summer and winter schedule, new requests for contracted capacity is submitted by the worldwide sales organization. Customers request capacity per flight/weekday. KLM determines for each flight/weekday how much of their total forecasted flight capacity may be sold on a contracted basis (on all flights to and from Amsterdam). Requests are then ranked per flight/weekday in descending orders beginning with the highest margin per cubic meter or per kilogram. Then final contracts are signed (KLM may take some lower margins over higher margins under certain considerations from their customers, particularly if their historical performance is reliable). Contract requests may also be submitted during the season, and go through a similar process. • MM on R/R On all flights to and from Amsterdam, route managers set and adjust shipment entry conditions (or SECs) daily. Given a daily SEC, minimum margin requirements are
defined. The SEC can be set either per cubic meter or per kilometer, depending if the flight is volume or payload constrained. Daily forecasts for SECs are made depending upon historic booking profiles, current capacity, current booking profile, market knowledge, and other considerations. Conference calls are held daily between route managers and the sales force to ensure SECs are raised or lowered appropriately. If a disruption causes SECs to be lowered unexpectedly (i.e., if there is a large unexpected cancellation), a flight alert message is sent to commercial organizations to attract more cargo to that flight. If the margin on the requested shipment exceeds (is less than) the SEC, the flight is accepted (rejected). If the booking involves multiple legs, the same principle is applied, now to the sum of the legs. The margin requirement is updated and incorporated into their information systems for transparency. KLM Cargo continues to improve the base of their margin management systems. They continually seek to improve upon their existing foundation by introducing new mechanisms so as to protect high margin cargo and eliminate low margin cargo, manage the relation between SEC and price, and to invest in upgraded systems to improve the robustness of their existing systems.
While not as prominent as with passenger security, cargo security has recently been under considerable scrutiny over its perceived vulnerability. This can be seen from the 108th Congress where some have called for screening of 100% of all cargo aboard passenger planes (see The Aviation and Transportation Security Act, P.L. 107-71), which nearly all experts in the industry will render as infeasible. However, there have been a number of policy changes recently calling for increased security measures. Among these include the TSA to at least tripling the amount of cargo placed on passenger aircraft that has been inspected (FY2005 Homeland Security Appropriations Act P.L. 108-334). Another calls for the TSA to pursue screening technologies and enhance security procedures as recommended by the 9/11 Commission (National Intelligence Reform Act of 2005, P.L. 108-458). Security concerns by no means unique; analogous reforms have been adopted in Germany, the U.K., Singapore, among other locations. While 100 percent of cargo is screened, far less than that is inspected. Screening refers exclusively to the known shipper program where physical inspection. Inspection is done a number of ways (credentialed agents, trained canines, biometric technologies, etc.). Here we seek to discuss issues pertinent to security, and the future of security design.
Air Cargo Security Risks
Let us first highlight the risks associated with air cargo security. 1. Explosive and Incendiary Devices – From the Congressional Research Service (CRS) Report to Congress in a mimeo updated January 26, 2006 (pp. 4-5): ”Undetected explosive or incendiary devices placed in air cargo are potential threats to aircraft. Experts have warned that air cargo may be a potential target for terrorists because screening and inspection of air cargo is currently not as extensive as required screening of passengers and checked baggage. Cargo carried aboard passenger aircraft may be at particular risk since passenger aircraft are generally regarded as highly attractive targets to terrorists and have been attacked in the past. It has been reported that TSA considers the likelihood of a terrorist bombing of a passenger airplane to be between 35% and 65% based on 2002 intelligence reports, and TSA believes that cargo is either likely to become, or already is, the primary aviation target for terrorists in the short term… While cargo as a means to place explosive or incendiary devices aboard aircraft has historically been rare, heightened screening of passengers, baggage, and aircraft may make cargo a more attractive means for terrorists to place these devices aboard aircraft, including all-cargo aircraft as well as passenger aircraft, in the future. Investigations have suggested that al Qaeda terrorists had an interest in bombing all-cargo aircraft prior to September 11, 2001, and were planning to bomb U.S.-bound cargo flights in an operation run out of the Philippines”
2. Hazardous Materials – about 75% of all hazardous materials shipped vis-à-vis aircraft are carried on cargo planes while the remaining 25% are shipped on passenger aircraft13. There are significant risks associated with shipping hazardous materials that are not declared leading to improper handling and storage. This can be seen from the May 11, 1996 crash of a ValuJet DC-9 over the Florida Everglades. The National Transportation Safety Board determined that improperly carried oxygen generators ignited in one of the plane’s cargo holds. 3. Cargo Crime – The U.S. General Accounting Office estimates that loses attributable to cargo theft across all transportation modes are between $10 and $25 billion in the U.S. alone, which a nontrivial share of this is from air freight. Crimes generally includes either theft of goods and the smuggling of contraband, counterfeit, and pirated goods through a distribution network. According to the CRS, there have been major theft rings at JFK Airport in New York, Logan Airport in Boston, and Miami International Airport. A recent review of transportation security needs has focused on six components to improve cargo security so as to abate cargo crime. 4. Aircraft Hijacking and Sabotage – Of course this area remains a great concern in a post 9/11 world. Many have argued that hijackings of all-cargo planes remain relatively more attractive given the recent improved security measures taken with respect to passenger aircraft. While historically rare, there have been some instances of hijacking of freighters.
B. Known Shipper Programs
One of the most important components to the security process is the so-called known shipper program. While this may be called something differently in other countries, the premise is that once a shipper has built a sufficiently high, cargo would be pre-screened and subject to lax security measures. The known shipper program has evolved as the principal means for pre-screening cargo. They were established initially to differentiate trusted shippers known to either a freight forwarder or an air carrier. Therefore containers from unknown shippers would undergo further screening and inspection. Currently in the U.S. cargo shipments from unknown shippers are prohibited on board passenger aircraft. The profiling procedure from the known shipper program is not that unlike passenger profiling, which is used in passenger security. The exact policies of the known shipper program vary across countries. Let us look at a few and how they differ. United States – Today there are more than 400,000 known shippers in the U.S. Most estimations show that between four and five percent of all cargo is inspected.
Source: U.S. General Accounting Office, “Aviation Safety: Undeclared Air Shipments of Dangerous Goods and DOT’s Enforcement Approach”, GAO 03-22, January, 2003.
There are generally two prominent criticisms of the U.S. system that are being focused on improving. The first, critics argue, is that the discretionary basis upon which shippers are classified as trustworthy. They argue that there is little oversight as forwarders and carriers have authority to accept/reject shippers into the program. The second criticism is the lack of a centralized database. Because of the preceding sentence, there are frictions for the TSA to monitor known shippers. The TSA has issued a proposed rule that would make consolidation of known shippers into a centralized database a requirement. Currently, shipments from unknown shippers are prohibited from being transported on passenger aircraft. Additionally, both carriers and forwarders must refuse to transport any cargo from shippers (irrespective of whether they are known or not) that refuse to give consent for searching and inspecting cargo. European Union – The most noteworthy observation about the EU regulated agent program is that all shipments who do not qualify as a “regulated agent” must be kept in storage by airfreight carriers for five days before they are eligible to be transported. This was implemented in 2006 and other countries are monitoring the results carefully. Moroever, there have been increased training requirements for security personnel as well as freight documentation. The latter poses challenges as firms are forced to modify IT systems accordingly. Hong Kong – In March 2000, Hong Kong implemented the Regulated Agent Regime, or RAR similar to the known shipper program in the U.S. While a non-regulated agent still has the ability to ship cargo subject to substantially higher security controls, these types of shipments constitute a negligible share of all cargo shipped from Hong Kong. At HACTL, over 95% of all cargo originating in Hong Kong is shipped from a registered agent. Regulated agents are likely to have their status stripped if either there are discrepancies on the air waybill, or more prominently if dangerous or hazardous materials are mislabeled or mishandled.
C. Air Cargo Security Technology
Given the joint increase in both the (expected) volume of air cargo flown and security concerns, a greater share of cargo can be inspected only when there are sufficient technological advances. Here we describe current technologies implemented in air cargo security and where the technologies are moving toward to accommodate future security. The National Intelligence Reform Act of 2004 (P.L. 108-458) has directed the TSA to develop technologies of air cargo security by authorizing $100 million annually from 2005-2007 for research and development of upgrading the existing security systems.
A number of technological advancements have already been made, yet still await implementation in air cargo security. Each of these was reported by the CRC (see Elias 2006) • Tamper-Evident and Tamper-Resistant Seals Various tamper-resistant technologies already exist in cargo shipments, but advancements have recently been made in electronic seals that would allow more immediate detection of tampering. • X-Ray Screening The most commonly used systems in the screening of cargo is x-tray technologies. One of the focal points moving forward is advancing current technologies that minimize human factors. To this extent, threat image projections (which superimpose images on x-ray scans) are one of the more promising advancements made. • Explosive Detection Systems Current explosive detection systems exist that use x-ray computed tomography in scanning objects, along with computational algorithms that estimate the probability of a threat given the density and shape of the package. Advancements are being made to include the possibility of scanning large objects (currently there are limitations as to the size of objects that can be scanned). • Chemical Trace Detection Systems These mechanisms are used in passenger screening, but are far less common in screening cargo considering these systems are labor intensive and time consuming. TSA has been contemplating extending the technology currently implemented in passenger screening to cargo screening. Given the labor input, this would probably be a secondary source of screening since it would be impractical to search a significant share. Nevertheless, this is under consideration and is less developed. • Neutron Beam Technologies While still being developed, neutron beam technologies remain one of the more exciting advances in security. These systems send a pulse neutron generator to an enclosed object that initiates multiple low energy nuclear reactions with the chemical elements comprising the object. Then the detector can measure this reaction and the gamma-rays emitted from the reaction, and information can be inferred from these measurements. While a promising means of securing the air cargo process, there are substantial costs involved in developing such methods. It is estimated from the General Accounting Office that with currently available neutron-based technology, each machine would cost about $10 million and would require about one hour per container. • Hardened Cargo Containers This is an attractive means to abate concerns about explosive devices. The 9/11 Commission suggested increasing the usage of hardened containers. Hardened
containers became popular after the 1988 bombing of Pan Am flight 103 over Lockerbie, Scotland. There currently exists hardened unit loading devices, although carrying these may induce inefficiencies from their increased weight which reduces available payload, as well as increased procurement cost for hardened containers. Therefore, current challenges are to strengthen the durability of these containers while introducing lighter material to offset these inefficiencies. The airlines are likely to impose a law requiring substantial increases in the use of hardened containers because of the crowd out effect of capacity. However, there are interesting questions to be raised, such as trading off lower capacity for lower security costs of unknown shippers (since their goods would be placed in hardened containers, the need to as stringent security would not be as high). • Biometric Screening Technology In order to ameliorate the problems associated with air cargo security risks, as in section VI.A (much of which has been caused by credentialed individuals with access to cargo loading areas), biometric screening technologies have been advanced. A pilot program has been established at some selected facilities whereby fingerprinting, retinal scans, and facial pattern recognition are being evaluated of transportation workers. The National Intelligence Reform Act of 2004 contains extensive provisions requiring the TSA to enhance the usage and implementation of biometric systems at airports (P.L. 108-458).
VII. Concluding Remarks and Future Work
Even in the presence of substantially higher expectations in oil prices, the air cargo industry still is expected to grow at a sufficiently high rate so as to induce more pressure in an already highly-capacitated system. Consequently, there are a number of areas for work to accommodate this growth, or at least to develop a greater understanding certain areas. I suggest four principal areas that seem particularly attractive. (i) The air cargo network is a very complex one with the number of possible routings, schedules, consolidation points, and stations serviced too many to enumerate. Fleet assignment, crew scheduling, recovery, operations, and revenue management have shaped the broad field known as airline operations research. However this usually refers to passenger movement, and less is known about cargo. The cargo model is vastly different than its passenger counterpart, and it could be argued that its network is even more complex than with passengers considering cargo has more routing options (unlike passengers, cargo that is not of highest priority could receive multiple transfers). Studying the planning process of the cargo network should be better understood as a first step to optimization applications. (ii) Continue the development of air cargo revenue management in three aspects. First, develop better forecasting methods or design incentives so as to reduce the uncertainty of show-up rates. Second, it would be of great interest to improve upon the existing revenue management principles so as to better ascertain the true marginal value of a booking request. As mentioned, modeling such phenomena is complex considering the dimensionality of the problem. Third, it is of interest to continue to foster the implementation of revenue management similar to what KLM has done. (iii) Develop a greater understanding and mechanisms for collaboration along the air cargo supply chain. The way in which to design an incentive compatible contract both among different players, as well as different firms as the same player would likely yield better network management and reduced lead times. (iv) Design more efficient warehouse and ground handling systems in an increasingly automated fashion. This is particularly motivated by major facilities facing severe capacity constraints (like LAX). As mentioned previously, this document sought three objectives. This was not meant to expand the frontier of academic literature, neither was it intended to make any proposals regarding the movement of air freight. Rather, we sought to obtain answers to the most salient questions such as what is likely to be shipped, how volumes of cargo have evolved and where they are anticipated to evolve, how exactly the process works, how different players along the supply chain interact and collaborate, identified the biggest firms in the industry and how they operate, and to understand issues that ought to be improved in shaping the air cargo industry of tomorrow.
Amaruchkul, K., Cooper, W.L., and Gupta, D. (2005) “Single-Leg Air-Cargo Revenue Management” Bazaraa, M., Hurley, Joseph D., et. al.,. “The Asia Pacific Air Cargo System”, mimeo, The Logistics Institute – Asia Pacific, 2000 Billings, J., Diener, A., and Yuen, B. (2003) “Cargo Revenue Optimisation”, Journal of Revenue and Pricing Management, vol. 2 pp. 69-79 Elias, Bartholomew, “Air Cargo Security”, CRS Report to Congress, version: January 26, 2006, The Library of Congress Kasilingham, R. (1996) “Air Cargo Revenue Management: Characteristics and Complexitites”, European Journal of Operations Research, vol. 96, pp. 36-44 Kleywegt, A.J. and Papastavrou, J.D. (2001) “The Dynamic and Stochastic Knapsack Problem with Random Sized Items”, Operations Research, vol. 96, pp. 26-41 Kleywegt, A.J., “Transportation and Supply Chain Systems: Air Cargo Transportation”, PowerPoint slides, accessed: www2.isye.gatech.edu/people/faculty/Alan_Erera/courses/air.pdf Pak, K. and Dekker, R. (2004) “Cargo Revenue Management: Bid-prices for a 0-1 Multi Knapsack Problem”, ERIM Report Series Research in Management, Rotterdam School of Management, Eerasmus Universiteit Rotterdam, The Netherlands Popescu, A. (2007) “Air Cargo Capacity and Revenue Management”, Doctoral Dissertation (forthcoming), H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology Sander, Charles “Initiatives in Aviation Procedure”, mimeo, Unisys Global Transportation, 2004. Accessed at unisys.com/transportation Slager, B. and L. Kapteijns (2004) “Implementation of Cargo Revenue Management at KLM”, Journal of Revenue and Pricing Management, vol. 3, pp. 80-90 Zondag, Willem-Jan (2006) “Competing for Air Cargo”, Master’s Thesis, Department of Spatial Economics, Free University Amsterdam Boeing World Air Cargo Forecast, 2006-2007, accessed at boeing.com/commercial/cargo/wacf.pdf Airbus Global Market Forecast, 2006-2025, accessed at
airbus.com/store/mm_repository/pdf/att00008552/media_object_file_AirbusGMF200 6-2025.pdf Lufthansa Cargo Security Dossier, “Air Cargo Security – Investing in the Future” Government Accounting Office, “DOT’s Efforts to Promote U.S. Air Cargo Carriers’ Interests”, October 1996, accessed at: www.gao.gov/archive/1997/rc97013.pdf Airforwarders Association “Industry Facts”, unknown date, accessed at www.airforwarders.org/industry_facts.php