Productivity Trends in the US Passenger Airline Industry 1978 ... - MIT [PDF]

Productivity Trends in the US Passenger Airline Industry 1978-2010 Peter Belobaba, Kari Hernandez, Joe Jenkins, Robert Powell, William Swelbar August, 2011

1

Foreword A Nation’s economic prosperity depends heavily on the efficiency and effectiveness of its transportation infrastructure. Thirty years have passed since the deregulation of the United States’ transportation industries, which have made notable changes in their operations. The purpose of Transportation@MIT’s US Transportation Productivity Study is to analyze the productivity trends and data for the transportation industries to understand what improvements have occurred in the past thirty years and why. In the 2010-2011 academic year, our team completed two studies, one on freight transported by rail and the other on passengers transported by air. Additional modes of transportation, including air freight, trucking and passenger rail, will be added to the study in 2011-2012. Beyond these industry studies, we plan to develop recommendations for government policy and investments, and for industry operations, on how to continue these productivity improvements and enhance our economic growth. This study was made possible by the generous funding from The Speedwell Foundation and The Shelter Hill Foundation. We would like to acknowledge Michael Messner and Paul Shiverick for their insightful feedback and support during our research and their engagement with faculty, research staff and students from the School of Engineering, the MIT Sloan School of Management and the School of Architecture and Planning.     Stephen  C.  Graves   Abraham  J.  Siegel  Professor  of  Management  Science   Professor  of  Mechanical  Engineering  and  Engineering  Systems       Rebecca  Cassler  Fearing   Executive  Director  of  the  Transportation@MIT  Initiative  

2

Contents EXECUTIVE SUMMARY  ...............................................................................................................................  5   1.0   Introduction  ...............................................................................................................................................  9   1.1   US  Airline  Industry  Overview  ........................................................................................................................  9   1.2   Evolution  of  the  US  Airline  Industry  .........................................................................................................  10   2.0   Trends in Traffic, Output, Fares and Profitability  .........................................................................  14   2.1   Growth  in  Passenger  Traffic  and  Airline  Output  ..................................................................................  14   2.2   Growth  of  Low  Cost  Carriers  ........................................................................................................................  17   2.3          Average  Fares  and  Total  Revenues  ...........................................................................................................  19   2.4   Industry  Profit  Volatility  ...............................................................................................................................  21   3.1   Airline  Operating  Cost  Categories  .............................................................................................................  23   3.2          Unit  Operating  Costs  per  Available  Seat  Mile  (CASM)  ........................................................................  25   3.3          Convergence  of  NLC  and  LCC  Operating  Costs  .......................................................................................  28   3.4   Distribution  Cost  Reductions  ......................................................................................................................  33   4.0   Trends in US Airline Productivity  ......................................................................................................  36   4.1     Airline  Inputs  and  Outputs  ..........................................................................................................................  36   4.2   Aggregate  Measures  of  Productivity  .........................................................................................................  36   4.3   Aircraft  Productivity  ......................................................................................................................................  38   4.4   Fuel  Efficiency  Measures  ...............................................................................................................................  41   4.5   Labor  Productivity  ..........................................................................................................................................  44   4.6     Multi-­‐Factor  Productivity  ............................................................................................................................  46   5.0   Evolution of Airline Networks and Airport Connectivity  ..............................................................  51   5.1   Hub-­‐and-­‐Spoke  versus  Point-­‐to-­‐Point  Networks  .................................................................................  51   5.2   US  Airport  Connectivity:  A  Passenger  Perspective  Since  1980  ........................................................  53   5.3   US  Airline  Domestic  Network  Development  Since  2000  ....................................................................  60   5.4   Summary  ............................................................................................................................................................  68   6.0   Looking Ahead: US Airline Industry Challenges  ..........................................................................  69   APPENDIX  ........................................................................................................................................................  74   A.1   Hub-­‐and-­‐Spoke  Analysis  Notes  (Hernández  et  al,  2011)  ...................................................................  74   A.2   Airport  Connectivity  Study  Notes  (Jenkins,  2011)  ...............................................................................  75  

      3

List of Figures and Tables Figure  2.1  :  Annual  US  Domestic  Revenue  Passenger  Miles  ..............................................................................................................................  14   Figure  2.2:  US  Airline  Output  ..........................................................................................................................................................................................  15   Figure  2.3:  Traffic  (RPMs),  Output  (ASMs),  and  Average  Load  Factor  ........................................................................................................  16   Figure  2.4:  Domestic  ASMs  by  Carrier  Type  .............................................................................................................................................................  17   Table  2.1:  Comparison  of  LCC  Business  Models  (Adapted  from  Belobaba  et  al,  2009)  .........................................................................  18   Figure  2.6:  Inflation  Adjusted  Total  Passenger  Revenues  by  Region  ............................................................................................................  20   Figure  2.7:  Inflation  Adjusted  US  Airline  Industry  Net  Profits  .........................................................................................................................  21   Figure  3.1:  Inflation  Adjusted  Total  US  Airline  Operating  Expenses  .............................................................................................................  24   Figure  3.2:  Major  Components  of  Unit  Costs  (Tsoukalas  et  al.  2008)  ...........................................................................................................  26   Figure  3.3:  Inflation  Adjusted  Unit  Costs  ...................................................................................................................................................................  26   Figure  3.4:  Inflation  Adjusted  Unit  Costs  by  Category  .........................................................................................................................................  27   Figure  3.5:  Inflation  Adjusted  Unit  Costs  (excl  Transport  Related  Expenses)  for  Domestic  Operations,  NLC  v.  LCC  ...............  28   Figure  3.6:  Inflation  Adjusted  Fuel  Unit  Cost  for  Domestic  Operations,  NLC  v.  LCC  ...............................................................................  29   Figure  3.7:  Inflation  Adjusted  Non-­‐Labor  Unit  Costs  for  Domestic  Operations,  NLC  v.  LCC  ...............................................................  30   Figure  3.8:  Inflation  Adjusted  Labor  Unit  Cost  for  Domestic  Operations,  NLC  v.  LCC  ............................................................................  32   Figure  3.9:  Inflation  Adjusted  Labor  Unit  Cost  for  Domestic  Operations,  NLC  v.  LCC  v.  Southwest  ................................................  33   Figure  3.11:  US  Airline  Commission  and  Sales  Costs  as  a  Percentage  of  Passenger  Revenue  ............................................................  35   Figure  4.2:  Aircraft  Utilization,  NLC  v.  LCC  ..............................................................................................................................................................  39   Figure  4.3:  Aircraft  Productivity  by  Carrier  Type  for  Domestic  Operations  ..............................................................................................  40   Figure  4.6:  Labor  Force  Productivity  and  Total  Labor  Force  ...........................................................................................................................  45   Figure  4.7:  Year-­‐by-­‐Year  Change  in  Total  MFP  Since  1980  ...............................................................................................................................  49   Figure  4.8:  Cumulative  Growth  of  MFP  Since  1980  ...............................................................................................................................................  50   Figure  5.1:  Example  of  Hub  Network  (Source:  Belobaba  et  al,  2009)  ..........................................................................................................  52   Figure  5.2:  Passenger  Weighted  Path  Quality  .........................................................................................................................................................  55   Figure  5.3:  Destinations  Flown  ......................................................................................................................................................................................  56   Figure  5.4:  Passenger  Weighted  Circuity  ...................................................................................................................................................................  57   Table  5.1:  Average  Excess  Miles  Flown  .......................................................................................................................................................................  58   Figure  5.5:  Inflation  Adjusted  Average  Passenger  Fares  ....................................................................................................................................  59   Figure  5.6:  Mainline  Domestic  Flight  Volumes  by  Carrier  .................................................................................................................................  60   Table  5.2:  US  Airline  Domestic  Hub  Operations:  LCC  v.  NLC  (Weighted  Average)  ..................................................................................  61   Figure  5.7:  Domestic  Mainline  Percentage  by  Carrier  .........................................................................................................................................  62   Figure  5.8:  PHL  Key  Carriers  ...........................................................................................................................................................................................  64   Figure  5.9:  PHL  Hub  Carriers  ..........................................................................................................................................................................................  64   Figure  5.10:  US  Airways,  America  West  Domestic  Mainland  Flight  VolumesFiFigure  ..........................................................................  66   Figure  5.11:  Delta,  Northwest  Domestic  Mainline  Flight  Volumes  .................................................................................................................  66   Figure  5.12:  United,  Continental  Domestic  Mainline  Flight  Volumes  ...........................................................................................................  67  

4

EXECUTIVE SUMMARY The Airline Deregulation Act of 1978 started a process of transformation for the US passenger airline industry that has accelerated in recent years. Over the past 30 years, air travelers have seen dramatically lower airfares as well as changes to route networks and service quality. At the same time, airlines experienced greater profit volatility and, in some cases, bankruptcy and liquidation. The airlines that survived this transition have also become significantly more cost efficient and productive. This report summarizes the findings of several studies of cost and productivity trends in the US passenger airline industry, undertaken at MIT during the past year. Productivity improvements during the first 50 years of airline industry development were driven primarily by innovations in aircraft technology. Until 1978, the ability of US airlines to achieve greater levels of productivity was constrained by economic regulation. It has only been in the period since deregulation that airlines have focused on cost efficiency and productivity improvement in the face of increasing competition. Traffic, Output, Fares, Revenues US domestic passenger air traffic, measured in revenue passenger miles (RPMs), has almost tripled since deregulation. The output of US passenger airlines, measured in available seat miles (ASMs), increased by 186% between 1978 and its peak output level of 2007. The mix of international versus domestic capacity has also changed – in 1978, only 19% of US passenger airline output was flown on international routes, whereas this proportion has increased to almost 30%. With RPMs increasing at a faster rate than ASMs, the average system load factor (percentage of available seats sold) has increased steadily since 1980. By 2009, average load factors for US airlines surpassed 80%, more than 20 percentage points higher than in the early 80s, with a large portion of the increase realized in the last decade. Higher load factors reflect improvements in productivity attributable to improved scheduling and fleet assignment practices and the development of differential pricing and revenue management techniques. While US domestic air travel has grown at rates significantly greater than prior to deregulation, average real fares have declined significantly and in 2009 remained at less than 50% of 1978 levels. Total industry passenger revenues have risen over much of the period as a result of increased output, traffic, and load factors, but have stagnated in the last decade – total 2009 passenger revenues were in real terms equal to 1988 levels, reversing 20 years of revenue growth. Increased competition led to an increase in the volatility of US airline profitability, as the total net profits of US airlines have been both cyclical and increasingly variable over the past 30 years. After the industry posted five consecutive years of losses from 1990 to 1994, it returned to record profitability in the late 1990s, before once again plunging into financial crisis and record operating losses between 2000 and 2005. 5

These aggregate measures illustrate the stark contrasts of US airline industry performance since deregulation. Traffic has tripled, while total output has increased by a slightly smaller amount, contributing to higher average load factors. Average real fares have decreased by about 50%, more so in US domestic markets. Total industry passenger revenues have grown more slowly due to lower fares, and recent competitive and economic impacts have reduced passenger revenues to 1988 levels in real terms. Overall, the US airlines have experienced stretches of record profitability since 1978, followed by stretches of even greater record losses, and they remain in a financially fragile condition. Operating Costs Total operating expenses in real terms (2010 dollars) have increased from $67 billion in 1978 to $108 billion in 2010, or by 61%. Compared to the 186% overall growth in ASM capacity over the same period, this relatively modest increase in real operating expenses suggests that significant improvements in cost efficiency have been achieved. Historically, fuel has accounted for a smaller portion of total operating expenses than in the most recent decade, peaking at over 36% of total airline operating expenses in 2008. Labor costs, on the other hand, have decreased substantially, especially since the re-structuring by Network Legacy Carriers (NLCs) in the early 2000s. The share of total operating expenses related to labor decreased from 42% in 1978 to 29% in 2010. Unit cost is the ratio of airline total operating expenses to ASMs produced, also known as CASM (cost per ASM). The average unit cost of US passenger airlines in real terms has declined almost 40% since deregulation. The largest portion of this decrease in unit operating costs has occurred in the labor cost category. Labor unit costs fell quickly in the early 1980s and then remained relatively stable until the early 2000s, when NLC re-structuring led to a more dramatic drop in labor costs. In real terms, labor unit costs have decreased by 55% since deregulation. The most volatile component of airline unit costs is fuel. Very high fuel costs in the early 1980s exceeded the recent peak in 2008, but much of the period from the late 1980s through the early 2000s was characterized by fairly low and stable real fuel unit costs. Low Cost Carriers (LCCs) have driven significant change in the US airline industry, with lower cost structures and higher productivity levels that allow them to offer lower fares and operate profitably. Although LCCs have historically reported unit costs about 2¢ per ASM lower than NLCs, unit costs have been converging for the two groups, particularly in recent years. LCC unit costs relative to NLCs were about 20% lower in 2009 compared to 30% lower in 2001, with this convergence explained largely by decreased labor unit costs of the NLC group. Productivity Trends Productivity is typically measured as the amount of output created per unit of input. In the airline industry, output is measured primarily as the capacity produced, or ASMs. RPMs are also used in productivity metrics as it can be argued that the ability of an airline to fill its ASMs with traffic (RPMs) captures additional facets of efficiency and productivity.

6

At the most aggregate level for the US passenger airline industry, total ASMs per real dollar of operating expense has increased by over 50% since 1978. In the early 1980s, US airlines produced just over 5 ASMs for each dollar (in real 2010 terms) of operating expense. In the most recent decade, productivity has increased to nearly 9 ASMs per dollar. Airline productivity can then be broken down according to key inputs: capital (aircraft), fuel, and labor. Aircraft utilization has increased dramatically since deregulation, peaking in 2007. LCCs have historically posted utilization about 1.5 block hours per day higher than NLCs. NLC aircraft utilization increased by about 10%, while LCC aircraft utilization increased by over 30% through 2007. Fuel productivity has increased dramatically since 1978: by 73% for produced output (ASMs) per gallon of fuel and by 128% for consumed output (RPMs) per gallon of fuel. By 2010, US airlines delivered 64 ASMs per gallon of fuel (and 52 RPMs), meaning the industry’s fuel efficiency exceeds that of the average automobile. The total number of employees in the US airline industry grew along with increasing capacity and traffic through the 1980s and 1990s, with temporary declines during economic downturns. After peaking at almost 550,000 in 2000, total US airline employment has plummeted by over 30%, due largely to the NLC labor force cuts in the early to mid-2000s. Overall, labor productivity has grown in waves since deregulation, and has reached historically high levels. ASMs per employee have more than doubled, increasing 108% since 1978, with more than half of that gain achieved since 2001. Preliminary estimates of the overall increase in multi-factor productivity (MFP) for the US passenger airline industry also show tremendous productivity improvement over the past 30 years. Use of the growth accounting methodology for the period since deregulation indicates that, on the basis of ASMs as output, aggregate airline MFP has increased by about 80%. Use of the RPM measure as output increases the estimate of MFP growth since 1980 to 160%, that is, the aggregate MFP of the US passenger airlines has grown by over 2.5 times when increases in average load factor are included. Airport Connectivity and Recent Network Evolution Hub-and-spoke networks allow airlines to provide joint supply of seats to multiple origindestination (O-D) markets with fewer flight departures and fewer aircraft, with lower total operating costs than in a point-to-point route network. Despite repeated forecasts of more pointto-point operations, the trend toward development of bigger and stronger hubs has continued. The economic advantages of hub network operations – increased revenues from more frequent (connecting) flight departures combined with the clear unit operating cost savings from operating fewer (and larger) aircraft than in a complete point-to-point network – far exceed their disadvantages. The US airline industry’s dependence on the hub-and-spoke model has continued to increase, with all NLCs reaching unprecedented levels of hub flights – well over 90% of their total operations. Even most LCCs, incorrectly thought to be point-to-point carriers, utilize a designated hub for over 90% of all flight segments. Only Southwest shows relatively low levels 7

of hub dependence, but even its use of hubbing has grown significantly, with over 50% of its flights arriving or departing a designated hub.

The effect of these changes in network structure on passengers has been lower average path quality and slightly higher circuity – both suggesting increased inconvenience of travel for nearly all airport categories studied. However, the apparent declines in these aggregate measures are due in part to the fact that far more passengers are choosing to select a connecting itinerary based on a lower fare. Improved airline efficiency from hubbing has lowered unit operating costs, and increased competition has forced the airlines to pass some of that cost savings to consumers in the form of lower fares. US Airline Industry Challenges The US passenger airline industry has undergone tremendous change since deregulation, with many of the most important changes to business practices, cost efficiency and productivity summarized in this report. Consumers have benefited from increased competition, lower fares, new entry and innovative service options, but airlines have not been able to retain the financial benefits from the many cost and productivity efficiencies they achieved. Despite all of the efforts of US carriers to restructure themselves in recent years, the industry remains in a vulnerable financial position. Moreover, the remnants of 60 years of regulation continue to affect the evolution of the US airline industry. Looking ahead for US airlines, global rather than domestic competition will shape of the future industry. The historical leadership of US airlines has been eroded, particularly during the most recent decade. While the US industry stagnated as it focused on the restructuring of costs and productivity, airlines in other regions of the world have continued to grow and have remained profitable. In the US industry, the NLC and LCC operating models have been converging, and with recent consolidation the two types of carrier will continue to co-exist. Airport and airspace infrastructure capacity constraints, along with the costs of expanding this infrastructure, are critical problems for the future of the airline industry. Although the FAA has been working toward increasing the capacity of the en route airspace, increasing congestion and delays indicate that the US air traffic infrastructure has not kept pace with air travel demand. Without major investments in new technologies and even additional airport infrastructure, it will be extremely difficult to accommodate the expected growth in air traffic. This study of US passenger airline productivity makes it clear that the airline business is both capital-intensive and labor-intensive, and is subject to a tremendous cyclicality driven primarily by economic forces and volatile fuel prices. Repeated cycles of record profitability followed by huge losses have left many US airlines in a weakened financial situation. Given that many elements of this cyclicality are not likely to change, the greatest challenge to the airline industry is to achieve sustained profitability and greater stability.

8

1.0

Introduction  

The passage of the Airline Deregulation Act of 1978 started a dramatic transformation of the US passenger airline industry that is not yet complete. With increased competition, air travelers have seen dramatically lower airfares as well as changes to route networks and service quality. At the same time, airlines have experienced increased profit volatility and, in some cases, bankruptcy and liquidation. Those that survived this transition have also become significantly more cost efficient and productive. The focus of this study is on the changes in operating efficiency and overall productivity of the US passenger airlines over the past three decades. This report summarizes the findings of several studies of cost and productivity trends in the US passenger airline industry, undertaken at MIT during the past year. A brief overview of the industry is provided in the remainder of this first section. Section 2 reviews changes in passenger traffic, average fares, industry output and profitability since 1978. In Section 3, the evolution of airline operating costs – fuel, labor, and non-labor components – is presented, with a focus on differences between established Network Legacy Carriers (NLCs) and new entrant Low Cost Carriers (LCCs). Section 4 examines aircraft and labor productivity measures in more detail, and presents the preliminary results of a Multi-Factor Productivity (MFP) study of the industry in aggregate. Section 5 summarizes the findings of two related studies – one looking at the changes in airport connectivity and accessibility from the passengers’ perspective, the other focusing on the recent evolution of US domestic airline networks. Finally, Section 6 looks ahead to the challenges that continue to face US airlines after more than 30 years of deregulation.

1.1

US Airline Industry Overview

In the US airline industry, approximately 100 certificated passenger airlines operate close to 10 million flight departures per year, and carry about one-third of the world’s total air passengers. US airlines enplaned 720 million passengers in 2010, 630 million of whom flew domestically. In 2010, US airlines (both cargo and passenger) generated $1.225 trillion in total US economic activity, contributing $731 billion – or 5.2% – of the US GDP, and provided 10.9 million jobs (ATA, 2011). 9

The US airline industry contributes significantly to both the US and global economies. Its economic impacts include direct effects like airline employment and many indirect effects on related activities that include aircraft manufacturers, airports, and tourism. The economic importance of the airline industry and its impacts on so many other major industries makes the volatility of airline profits and the financial sustainability of airlines a national concern. Yes, despite being critical to the nation’s economic activity, the US airline industry remains a target of regulation and taxation. Today, the US airline industry, its passengers and cargo are subject to 17 different federal taxes totaling nearly $17 billion per year (Calio, 2011), compared to an inflation-adjusted total of $6.2 billion twenty years ago. Thus, while average fares have decreased and total US airline passenger revenues have grown by only 7% in real terms, total taxes collected have increased by 174%. Based on the average price of a one-way domestic ticket, approximately $50 or more than 16 percent of the passenger fare is some form of tax (Karlsson, 2010). Continued taxation and emerging regulations being promoted in the name of consumer protection are imposing costs on the industry that could lead to unintended consequences – a smaller industry contributing less in economic activity than it does today.

1.2

Evolution of the US Airline Industry

Productivity improvements during the first 50 years of the US passenger airline industry were driven primarily by innovations in aircraft technology related to speed and capacity – the introduction of jet airplanes in the 1960s, followed by wide-body “jumbo jets” in the 1970s. Yet, until 1978, the ability of US airlines to achieve greater levels of productivity was constrained by economic regulation – management decisions as to which routes to serve, how often and at what price were subject to government controls. It has only been since deregulation that airlines have been able to focus on cost efficiency and productivity improvement in the face of increasing competition. Airline deregulation has benefited the vast majority of air travelers in the United States. Domestic air travel in the past 30 years has grown at rates significantly greater than prior to deregulation, while average real fares have dropped dramatically and in 2010 are still about onehalf of 1978 levels. New entry by innovative Low Cost Carriers (LCCs) contributed to increased fare competition, forcing the more established Network Legacy Carriers (NLCs) to reduce costs 10

and improve productivity, and changing the traveling public’s expectations with respect to lowpriced air travel. On the other hand, deregulation of US airlines also had some negative impacts. Cost cutting, increased profit volatility, mergers and bankruptcies of several large airlines, led to job losses and reduced wages for many airline employees. Residents of some small cities saw changes to their air services, as deregulated airlines were no longer obligated to serve less profitable routes with as much capacity or frequency. And, the development of large connecting hub networks by the NLCs also raised concerns about the pricing power of dominant airlines at hub airports (GAO, 1993). Deregulation removed barriers to entry into the US airline industry, spawning new entrant airlines with lower cost structures that allowed them to offer consumers new options for air travel at lower fares. Their pricing strategies, combined with the additional capacity offered in affected markets, reduced average fares for consumers and, in turn, the revenues of NLCs as they matched the lower prices to protect market share. The significantly lower cost structures of the LCCs allowed them to generate operating profits even at low fares, while the NLCs had little choice but to re-structure their operating models in the hopes of maintaining profitability. The significantly lower cost structures of the LCCs can be attributed primarily to higher levels of productivity of both aircraft and employees. LCCs initially operated “point-to-point” networks with simplified passenger processing and lower aircraft ground times, in contrast to the hub-andspoke networks of the NLCs. Shorter ground times enabled LCCs to achieve higher aircraft utilization rates than NLCs, contributing to lower unit aircraft operating costs. LCCs were also able to achieve significantly higher labor productivity than NLCs, due to more flexible work rules that allow cross-utilization of employees, which also contributed to lower unit labor costs (Belobaba et al, 2009). Although deregulation legislation was passed in 1978, it has taken several decades for its full impacts to be felt in the industry. In the years immediately following deregulation, a fuel crisis and economic recession clouded any assessment of its initial effects. While some new entrant LCCs began to emerge in the mid-1980s, existing NLCs were able to fend off the competitive

11

attacks through aggressive price matching. Barriers to entry had been removed, but there remained a variety of barriers to exit that allowed some inefficient legacy airlines to survive. The early 1990s brought the first Gulf War, a fuel crisis and economic recession that plunged the US airlines into another period of operating losses. Iconic names like Braniff, Eastern and Pan Am disappeared from the industry. By the mid 1990s, the remaining legacy airlines were able to return to profitability by reinforcing their hub and spoke networks, protecting their market share and keeping LCCs at bay. The “Big 6” NLCs were able to co-exist and generate record profits during the late 1990s, the last period of extended profitability for US airlines as a group. The financial problems of the US airlines that began with the economic downturn at the beginning of 2001 reached crisis levels between 2001 and 2005. The combination of the terror attacks of September 11, 2001, the subsequent economic downturn, several military actions, along with international health concerns drove the US airline industry into uncharted financial territory. Four out of the six US NLCs (US Airways, United, Delta and Northwest) filed Chapter 11 bankruptcy between 2001 and 2005. Under bankruptcy protection, these carriers focused on down-sizing, cutting operating costs and improving productivity as part of their re-structuring efforts. NLC airline employment dropped by 30% in just five years, representing over 100,000 jobs lost, while average wage rates were also cut by 7% (US DOT, 2011). Despite these restructuring efforts, US airlines posted cumulative net losses of over $60 billion from 2001 to 2010. In response to the challenges since 2000, the US airline industry achieved productivity gains that exceed the gains made during the first two decades of deregulation. Recent productivity gains have come from the introduction of new technologies (e.g., internet ticket distribution, web check-in) and by re-allocating capacity (e.g., moving aircraft from domestic to international routes in an effort to improve both aircraft and employee utilization). The NLCs have also attempted to replicate some of the cost efficiencies of the LCCs, for example, by eliminating free meals and pillows on domestic flights to reduce costs and by reducing aircraft turn-around times to improve aircraft productivity. Yet, three decades after deregulation, the industry remains fragile, with airlines struggling to find a business model that can ensure sustained profitability. Although fragile, the industry is 12

nonetheless much stronger today than it has been since 1978. Over the first 25 years of deregulation, the industry paid the equivalent of $30 per barrel for jet fuel (cost of crude oil plus the refining margin or “crack spread”). Today airlines are paying nearly $100 more per barrel for jet fuel and most of the airline companies are reporting profits, albeit modest profits. Without the restructuring that took place during the “lost decade” of the 2000s, it is doubtful that many of the iconic names would have survived as standalone companies after the most recent financial crisis. The business model that has been adopted by nearly every US airline following the schedule reductions between 2008 and 2010 is one of capacity discipline and focusing on profits instead of incremental revenue and/or market share. With this capacity discipline comes the capability to increase yields and revenues. Without it, airlines would not be able to pass through increasing proportions of the rising cost of jet fuel to the consumer. The increasing cost of oil has also encouraged consolidation as another approach to capacity discipline, and this recent consolidation has not been limited to only the NLCs but has affected the LCCs and regional sector as well.

13

2.0

Trends in Traffic, Output, Fares and Profitability

The evolution of the US passenger airline industry described in Section 1 is illustrated in this section with aggregate industry data for the period since deregulation. Changes in passenger traffic and industry output are presented, along with overall trends in load factors (percentage of seats filled), average fares and industry revenues. The growth of LCCs in the US domestic market is discussed, along with the increasing volatility of airline profitability.

2.1

Growth in Passenger Traffic and Airline Output

US domestic passenger air traffic, measured in revenue passenger miles (RPM), has almost tripled during the period 1978-2009, as shown in Figure 2.1. The annual growth of domestic passenger air traffic has been consistently positive, with only a few exceptions – traffic declined in 1980-1981 due to a recession, in 1991 due to the first Gulf War, the subsequent fuel crisis and economic recession, again in 2001-2002 after the 9/11 terror attacks; and most recently with the economic recession that followed the 2008 financial crisis. Figure 2.1 : Annual US Domestic Revenue Passenger Miles Figure  2.1:  Annual   US  Domestic  Revenue  Passenger  Miles 700

500

400

300

200

100

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Revenue  Passenger  Miles  (Billions)

600

14

The output of US passenger airlines, measured in available seat miles (ASMs), has also almost tripled over the same 30-year period. Figure 2.2 shows that total US passenger airline output, domestic and international combined, increased by 186% between 1978 and the peak output level of 2007. The increase in domestic ASMs is somewhat smaller, at 159%, reflecting the shift of capacity by most NLCs from domestic markets with high levels of LCC competition to international markets offering more favorable competitive conditions and operating cost economics. In 1978, only 19% of US passenger airline output was flown on international routes, whereas this proportion increased to almost 30% by 2009. Figure 2.2: US Airline Output Figure  2.2:  US  Airline  Output 1,200  

Available  Seat  Miles  (Billions)

1,000  

800  

600  

400  

200  

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

0  

Domestic

Atlantic

Latin  

Pacific

The combination of traffic (RPMs) and output (ASMs) determines the average system load factor, defined as RPMs divided by ASMs or, more simply, the proportion of airline output that is actually consumed. Figure 2.3 shows the overall growth in total system RPMs and ASMs for US passenger airlines, along with the resulting average load factors. With RPMs increasing at a faster rate than ASMs over much of the period, the average system load factor has increased steadily since 1980, with particularly dramatic increases since 2000. By 2009, average load 15

factors for US airlines surpassed 80%, more than 20 percentage points higher than in the early 80s, with a large portion of the increase realized in the last decade. This dramatic increase in average load factors, while perceived negatively by consumers hoping to sit next to an empty seat, represents a major improvement in airline productivity. The greater the proportion of airline output that is actually sold and generates revenues, the greater the potential for higher productivity and profitability. The increase in average load factors can be attributed to factors affecting both capacity (ASMs) and traffic (RPMs). On the capacity side, airlines have improved their scheduling and fleet assignment practices with more sophisticated techniques designed to allocate the right size of aircraft to individual routes during specific seasons, days of the week and times of the day. On the traffic side, the development of differential pricing schemes and revenue management techniques has allowed airlines to fill each flight departure with more revenue (and passengers). Figure 2.3: Traffic (RPMs), Output (ASMs), and Average Load Factor Figure  2.3:  Traffic  (RPMs),  Output  (ASMs),  and  Average  Load  Factor 85%

1,200  

80%

1,000  

Load  Factor

800   70% 600   65% 400  

RPMs/ASMs  (Billions)

75%

60% 200  

55%

50%

-­‐ 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Load  Factor

ASMs

RPMs

16

2.2

Growth of Low Cost Carriers

  Although several new entrant airlines emerged soon after deregulation, most failed during the 1980s and 1990s. The surviving low-fare airlines (also known as low-cost carriers or “LCCs”) grew during the 1980s, but still accounted for less than 7% of US domestic airline capacity (ASMs) in 1991. Their slow but steady growth continued until 2000, when the LCC sector began to grow more rapidly and to capture significant market share as the NLCs faced a financial crisis and cut domestic capacity. As shown in Figure 2.4, LCCs provided nearly 27% of domestic ASM capacity in 2009. In terms of total US airline output (including international ASMs), the LCC sector accounted for about 20% in 2009, as most LCCs have not expanded to many international markets. Also shown in Figure 2.4 is the growing proportion of domestic ASMs being provided by the “Express Carrier” (EC) section, namely regional airlines operating feeder services for the NLCs. Figure 2.4: Domestic ASMs by Carrier Type Figure  2.4:  Domestic  ASMs  by  Carrier  Type 100% 90%

70% 60%

50% 40% 30% 20%

10% 0%

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Share  of  Domestic  Output

80%

NLCs

LCCs

ECs

.

17

There are many characteristics and operating strategies commonly assumed to be shared by LCCs. These characteristics include use of a single aircraft type or interchangeable family of aircraft, “point-to-point” operations instead of connecting hub networks, no labor unions and lower wage rates for employees, single cabin service with no premium classes, no seat assignments, reduced on-board “frills” and seating space, no frequent flyer loyalty programs, and an avoidance of traditional distribution channels – all in pursuit of greater productivity and lower unit operating costs. However, many of these characteristics do not accurately represent the actual strategies employed by the largest and most successful low-cost carriers in the US. Table 2.1 presents a comparison of the three largest LCCs in the US, showing the extent to which each airline meets the above “typical” LCC characteristics: Table 2.1: Comparison of LCC Business Models (Adapted from Belobaba et al, 2009) Southwest

JetBlue

AirTran

Single aircraft type or single family of aircraft

Y

N

N

Point-to-point ticketing, no connecting hubs

N

N

N

No labor unions, lower wage rates

N

Y

N

Single cabin service, no premium class

Y

Y

N

No seat assignments

Y

N

N

Reduced frills for on-board service (vs. legacy)

N

N

N

No frequent-flyer loyalty program

N

N

N

Avoid global distribution service (GDS)

?

N

N

Southwest Airlines is the oldest LCC, yet it does not adhere to the strict “LCC recipe”. While the airline’s network is not a classical hub structure, it has many “focus cities” at which a large proportion of its passengers make connections. And, Southwest is the most heavily unionized airline in the US (Gittell, 2003). JetBlue, launched in 2000, has even fewer of the typical LCC characteristics than Southwest. It operates two different aircraft types and it offers advance seat assignments and on-board service that is perceived by consumers to be superior to that of legacy 18

airlines. AirTran, the third largest US LCC, has none of the typical characteristics of LCCs listed in Table 2.1, yet is a successful low-cost operator.

2.3

Average Fares and Total Revenues

While US domestic air traffic has almost tripled since deregulation, average real fares have declined significantly and in 2009 remained at less than 50% of 1978 levels. Figure 2.5 shows the dramatic decreases in yield (average fare paid per passenger-mile), inflation adjusted to 2010 dollars, for both domestic and international travel on US carriers. The decrease in real fares has been greater in the US domestic market: international average fares dropped 49% between 1977 and 2009, and domestic average fares decreased at an even higher rate of 56% during the same time period. Figure 2.5: US Airline Inflation Adjusted Average Yield Figure  2.5:  US  Airline  Inflation  Adjusted  Average  Yield $0.35 $0.30

$0.20 $0.15 $0.10 $0.05 $0.00 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Passenger  Fare  per  RPM

$0.25

Domestic

System

International

19

Despite the significant decline in real fares since deregulation, total industry passenger revenues have continued to rise over most of the period as a result of increased output, passenger traffic, and load factors. Figure 2.6 shows the total passenger revenues for the industry in 2010 dollars. Whereas both output and traffic almost tripled, traffic, total industry passenger revenues in real terms have not even doubled, given the significantly lower average fares. And, of greater concern to industry performance, total passenger revenues have stagnated during the last decade, with dramatic revenue losses during the re-structuring period 2001-2005, and again after the financial crisis of 2008. Total passenger revenues in 2009 for the US airline industry were equal in real terms to the industry total back in 1988, effectively wiping out 20 years of revenue growth. Figure 2.6: Inflation Adjusted Total Passenger Revenues by Region Figure  2.6:  Inflation  Adjusted  Total  Passenger  Revenues  by  Region $120

Total  Passenger  Revenue  (Billions)

$100

$80

$60

$40

$20

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0

Domestic

Atlantic

Latin  

Pacific

   

20

2.4

Industry Profit Volatility

  The increased competition of deregulation led to an increase in the volatility of US airline profitability. As shown in Figure 2.7, the total net profits of US airlines have been both cyclical and increasing variable over the past 30 years. After the US airline industry posted five consecutive years of losses totaling over $13 billion from 1990 to 1994, it returned to record profitability in the late 1990s, with total net profits in excess of $34 billion between 1995 and 2000. Then, the industry once again plunged into financial crisis and record operating losses between 2000 and 2005, returning to profitability in 2006. The losses in 2008 were once again record-setting, with another return to profitability in 2010. Figure 2.7: Inflation Adjusted US Airline Industry Net Profits Figure  2.7:  Inflation  Adjusted  US  Airline   Industry  Net  Profits $20  

$0  

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Pre-­‐Tax  Profit  (Billions)

$10  

($10)

($20)

($30)

($40)

These aggregate industry measures – traffic, output, fares, revenues and profitability – illustrate vividly the contrasts of US airline industry performance since deregulation. Traffic has tripled, while total output has increased by a slightly smaller amount, contributing to higher average load factors. Average real fares have decreased by about 50%, more so in US domestic markets. At 21

the same time, total industry passenger revenues have grown more slowly due to these lower fares, and recent competitive and economic impacts have reduced total industry passenger revenues to 1988 levels in real terms. Overall, the US airlines have experienced cyclical stretches of record profitability almost inevitably followed by stretches of even greater record losses.

22

3.0

Trends in US Airline Operating Costs  

The competition made possible by deregulation focused the attention of airlines on cost containment, particularly given competition from new entrant LCCs with lower cost structures. In this section, we examine the trends in US airline operating costs since 1978, with a focus on unit operating costs in three major categories – fuel, labor and non-labor costs. While the industry has made tremendous progress in terms of unit cost efficiencies in the labor and nonlabor categories, the instability of fuel costs has proven to be a driver of the profitability cycles discussed above. A comparison of trends in NLC and LCC unit costs is also presented here, highlighting the extent to which there has been cost convergence between the two groups of airlines, particularly in the past 10 years.

3.1

Airline Operating Cost Categories

An airline’s operating costs can be divided into three major categories: fuel, labor, and nonlabor expenses. Fuel expenses are the most straightforward to categorize, and changing fuel prices have historically been assumed to affect all airlines equally. However, differences in the fuel efficiency of airline fleets as well as the use of financial hedging instruments by some airlines can result in variations in reported fuel costs, among airlines as well as over time. Labor costs include total salaries, benefits and other costs paid by airlines to employees, providing an indication of the use and cost efficiency of labor inputs. Non-labor costs include all other operating expenses not included in the fuel or labor-related cost categories. This last category includes cost items that represent the “structural” costs of the airline over which management can exert influence and are therefore a good gauge of how management strategies affect “controllable costs” not related to fuel or labor inputs. Figure 3.1 shows the evolution of total airline operating expenses broken down into these cost components on an inflation-adjusted basis since 1978, with all other detailed expense items shown as “non-labor” costs. For the US airline industry overall, total operating expenses in real terms (2010 dollars) have increased from $67 billion in 1978 to $108 billion in 2010, or by 61%. Compared to the 186% overall growth in ASM capacity over the same period, this relatively modest increase in real operating expenses suggests that significant improvements in cost efficiency have been achieved. 23

Historically, fuel has accounted for a smaller portion of total operating expenses than in the most recent decade. Fuel costs reached a peak of over 36% of total airline operating expenses in 2008, compared to about 30% during the first fuel crisis after deregulation in 1980. The proportion of total operating expenses attributable to fuel returned to 29% in 2010, still greater than that observed during most of the 1980s and 1990s. Labor costs, on the other hand, have declined in terms of both their absolute and relative contribution to total operating expenses, especially since the NLC re-structuring of the early 2000s. The share of total operating expenses related to labor decreased from 42% in 1978 to 29% in 2010. Non-labor costs have fluctuated as a proportion of total expenses, as the contributions of many of the smaller components of non-labor costs have changed: Outside Maintenance, Non Aircraft Ownership, and Aircraft Rental costs have increased, while the once significant Commissions (travel agency payments) category has all but disappeared. Figure 3.1: Inflation Adjusted Total US Airline Operating Expenses Figure  3.1:  Inflation  Adjusted  Total  US  Airline  Operating  Expenses $140 $120

$80

$60

$40

$20

$0

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Total  Expense  (Billions)

$100

Fuel

Labor

Non-­‐Labor

24

3.2

Unit Operating Costs per Available Seat Mile (CASM) Total operating expenses have increased since 1978, even on an inflation-adjusted basis,

because US airlines have increased output and are carrying more traffic. Unit cost is the ratio of airline total operating expenses to ASMs produced. For passenger airlines, unit cost is also known as “CASM”, meaning “Cost per ASM”. Unlike total operating costs, the relationship between unit costs on the one hand and volume of airline output is expected to be negatively correlated. That is, larger airlines expect to see some economies of scale (reduction in unit costs with increased output), as fixed costs are spread over a larger output of ASMs. All else equal, larger capacity aircraft should show some economies of aircraft size, as fixed costs are spread over more seats for any given flight, resulting in lower unit costs. And, longer stage lengths mean that the relatively fixed costs ground servicing, for example, can be spread over more ASMs produced.

Total unit costs as reported by US airlines to the DOT Form 41 include a category called “transport related expenses” which consists largely of payments made by large airlines to regional carriers to provide connecting services on their behalf. These connecting services provide the paying airline (usually a NLC) with incremental connecting traffic and revenue. However, because these payments are not actual “operating expenses” incurred in the production of the ASM output of the mainline carrier, they should not be included for the purposes of unit cost comparisons across time or among airlines (Tsoukalas et al, 2008). This adjustment is especially important when comparing the unit costs of NLCs and LCCs, given that LCCs do not typically rely on regional partners for connecting traffic feed. Figure 3.2 illustrates the four reported components of unit costs. Transport-related expenses are excluded from all of the comparisons presented here. The remaining three cost components of interest are fuel expenses, labor costs and non-labor costs.

25

Figure 3.2: Major Components of Unit Costs (Tsoukalas et al. 2008)

Figure 3.3: Inflation Adjusted Unit Costs Figure  3.3:  Inflation  Adjusted  Unit  Costs $0.20

$0.10

$0.05

$0.00 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Cost  per  ASM

$0.15

Fuel

Labor

Non-­‐Labor

26

Using these cost categories and excluding Transport Related expenses, Figure 3.3 shows the aggregate industry unit cost of providing capacity, expressed in 2010 dollar terms. The average unit cost of US passenger airlines has declined almost 40% since deregulation. In 1979, it cost the average US airline an inflation adjusted 18.3¢ to produce one ASM; that unit cost dropped to 11.2¢ in 2009. Figure 3.4: Inflation Adjusted Unit Costs by Category Figure  3.4:  Inflation  Adjusted  Unit  Costs  by  Category $0.09

Inflation  Adjusted  Unit  Costs  by  Category

$0.08 $0.07 $0.06

$0.05 $0.04 $0.03 $0.02 $0.01

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0.00

Fuel

Labor

Non-­‐Labor

Figure 3.4 shows the inflation adjusted unit costs for the fuel, labor, and non-labor categories. Non-labor unit costs increased through the late 1980s and peaked in 1992. Since then, non-labor costs have fallen gradually, with the exceptions of 2001 and 2008, years in which output had to be reduced quickly in a response to reduced demand. Such rapid reductions in capacity cannot be matched with equally rapid reductions in the fixed costs that contribute to the non-labor cost category. A substantial part of the reduction in real non-labor unit costs since the mid-1990s can be attributed to substantial cuts in airline distribution costs – first with the elimination of travel agency commissions in the late 1990s, followed by the use of internet and related technologies 27

for ticket distribution since 2000. Overall, non-labor costs have decreased by 25% in real terms since 1978. Labor unit costs fell quickly in the early 1980s and then remained relatively stable until the turn of the century. A dramatic drop in labor unit costs occurred between 2002 and 2006, with NLC bankruptcies, layoffs and restructuring of labor contracts. In cumulative terms, the average real labor unit cost for US airlines has decreased by an astounding 55% since deregulation. Together, the labor and non-labor operating cost categories combined (excluding fuel) have seen a 40% decrease in real unit costs since 1978. Fuel unit costs expressed in inflation adjusted terms, on the other hand, have exhibited much greater volatility than the other two cost categories. Very high fuel unit costs in the early 1980s exceeded the recent peak in 2008 in real terms, but much of the period from the late 1980s through the early 2000s was characterized by fairly low and stable real fuel unit costs. Fuel unit costs began to surge in 2005, peaking in 2008 and leading to the US airline industry’s most recent profitability challenge.

3.3

Convergence of NLC and LCC Operating Costs

Figure 3.5 compares inflation adjusted NLC and LCC airline unit costs since 1978. The NLC group has reported total unit costs (excluding transport-related expenses) that have consistently been approximately two cents (USD) per ASM higher than the LCC aggregate. In percentage terms, the unit costs of the two airline groups appear to be converging. LCCs still had a clear unit cost advantage in 2009, but their unit costs relative to NLCs were about 20% lower in 2009 compared to 30% lower in 2001.

Figure 3.5: Inflation Adjusted Unit Costs (excl Transport Related Expenses) for Domestic Operations, NLC v. LCC 28

Figure  3.5:  Inflation  Adjusted  Unit  Costs  (excl  Transport  Related  Expenses)  for   Domestic  Operations,  NLC  v.  LCC $0.25

Cost  per  ASM

$0.20

$0.15

$0.10

$0.05

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0.00

NLCs

LCCs

As an exogenous input, fuel unit costs reflect market oil prices and have been very similar for NLCs and LCCs throughout the last three decades, as shown in Figure 3.6. Real fuel unit costs have fluctuated dramatically, but in 2010 were about 20% lower for both carrier groups than in 1978. To the extent that real unit costs in the fuel category differ between NLCs and LCCs, Figure 3.6 shows that NLCs have typically been more affected by fuel price increases, as NLCs have tended to operate older and less fuel-efficient aircraft than LCCs, on average. The 20032009 difference in fuel unit costs can be explained primarily by Southwest’s fuel hedging success that enabled the airline to weather the jump in oil prices more easily than the rest of the industry players who did not gamble similarly in the early 2000s.

Figure 3.6: Inflation Adjusted Fuel Unit Cost for Domestic Operations, NLC v. LCC 29

Figure  3.6:  Inflation  Adjusted  Fuel  Unit  Cost  for  Domestic  Operations,   NLC  v.  LCC $0.07 $0.06

Fuel  Cost  per  ASM

$0.05 $0.04 $0.03 $0.02 $0.01

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0.00

NLCs

LCCs

Figure 3.7 shows that the non-labor unit cost component has been declining slowly in real terms for both groups. The non-labor unit cost gap between the groups has remained approximately 1 cent per ASM. The non-labor CASM reflects the airline’s structural costs that are driven by a variety of factors such as network structure, fleet type, and outsourcing activity to name a few. This comparison of NLCs and LCCs suggests, therefore, that the NLCs have certain structural costs (hub operations, international flights, lounges and other services) that result in an inherent and consistent non-labor unit cost gap that appears to be about 1 cent per ASM.

Figure 3.7: Inflation Adjusted Non-Labor Unit Costs for Domestic Operations, NLC v. LCC 30

Figure  3.7:  Inflation  Adjusted  Non-­‐Labor  Unit  Costs  for  Domestic  Operations,   NLC  v.  LCC $0.09 $0.08

Non-­‐Labor  Cost  per  ASM

$0.07 $0.06 $0.05 $0.04 $0.03 $0.02

$0.01

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0.00

NLCs

LCCs

Labor unit costs, on the other hand, have changed dramatically in the most recent decade as shown in Figure 3.8. While labor costs in real terms have gradually declined for LCCs, NLCs experienced a dramatic downturn after 2002 as several of the largest carriers were able to renegotiate labor contracts and reduce workforces after filing Chapter 11 bankruptcy. This decline in labor unit costs led to a substantial narrowing of the historic gap between NLCs and LCCs, from an inflation adjusted1 1.8¢ per ASM in 1990 to 0.8¢ in 2009. NLC real labor unit costs dropped by 40% between 2002 and 2007, increasing thereafter as many of the labor contracts of the early 2000s were up for re-negotiation by 2008.

                                                                                                                1 All inflation adjusted values are presented in 2010 dollars. 31

Figure 3.8: Inflation Adjusted Labor Unit Cost for Domestic Operations, NLC v. LCC Figure  3.8:  Inflation  Adjusted  Labor  Unit  Cost  for  Domestic  Operations,   NLC  v.  LCC $0.12

Labor  Cost  per  ASM

$0.10

$0.08

$0.06

$0.04

$0.02

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0.00

NLCs

LCCs

To further examine the differences in labor unit costs, Figure 3.9 isolates Southwest from the remaining LCCs. As previously mentioned, Southwest has the most heavily unionized workforce in the industry and compensates its employees competitively with NLCs. Following the NLC restructuring, Southwest labor unit costs actually surpassed the average NLC labor unit costs for the first time in 2005. The labor unit cost gap between NLCs and all other LCCs remained significant, although narrowed from an inflation adjusted 2.9¢ per ASM in 1990 to 2.1¢ in 2009. The major factor driving total unit cost convergence between NLCs and LCCs is labor unit costs, as confirmed by Figures 3.7 and 3.8.

32

Figure 3.9: Inflation Adjusted Labor Unit Cost for Domestic Operations, NLC v. LCC v. Southwest

Figure  3.9:  Inflation  Adjusted  Labor  Unit  Cost  for  Domestic  Operations,   NLC  v.  LCC  v.  Southwest $0.12

Labor  Cost  per  ASM

$0.10

$0.08

$0.06

$0.04

$0.02

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

$0.00

Domestic

3.4

LCC  Domestic  excl  SW

SW

Distribution Cost Reductions One of the most significant cost reductions by US airlines since deregulation was

achieved in the sales and distribution of passenger tickets. With deregulation, the intense competition for passengers and market share extended to the relationship between airlines and their primary sales channels – travel agencies. By the early 1990s, it was not uncommon for airlines to pay travel agencies sales commissions of 10 percent of the fare for domestic tickets; 12 to 15 percent for transatlantic tickets; and over 30 percent of the fare of transpacific tickets. As shown on Figure 3.10, total US airline commission payments in real terms (2010 dollars) peaked at over $12 billion in 1993, increasing from just over $3 billion in 1978. As a percentage of passenger revenues, commission payments doubled from about 5% to almost 13% between 1978 and 1993.

33

Led by Delta Air Lines starting in 1994, the US airlines undertook a fundamental change to their sales practices by first reducing and ultimately ending the payment of commissions to travel agents for selling tickets.

By 2000, travel agency commissions had largely been

eliminated for US domestic tickets, and total annual commission costs for US airlines dropped by 41%, despite a 46% increase in total passenger revenues. As shown in Figure 3.10, the US industry achieved over $10 billion in annual savings by 2009, as total commissions fell to 1.4% of passenger revenue. Figure 3.10: Inflation Adjusted US Airline Total Commission Cost

14%

14  

12%

12  

10%

10  

8%

8  

6%

6  

4%

4  

2%

2  

0%

0  

Total  Commission  Cost  (Billions)

Commission  Cost  as  a  Percentage  of  Passenger  Revenue

Figure  3.10:  Inflation  Adjusted  US  Airline  Total  Commission  Cost

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Total  Commission  Cost

Percent  of  Passenger  Revenue

 

Another significant change to distribution practices was the development and acceptance of the Internet as a highly efficient sales channel for airlines. NLCs and LCCs alike took advantage of this rapidly changing technology after 2000, promoting direct distribution of tickets through airline web sites, the most cost-effective of all distribution channels. With the additional cost savings made possible by Internet distribution, total promotion and sales costs as a percentage of passenger revenues plummeted from their peak of over 22% in 1994 to 7.4% in 2008 (Figure 3.11). 34

Figure 3.11: US Airline Commission and Sales Costs as a Percentage of Passenger Revenue

Figure 3.11: US Airline Commission and Sales Costs as a Percentage of Passenger Revenue 25%

Percent  of  Passenger  Revenues

20%

15%

Commissions 10%

Sales  &  Distribution

5%

0%

The additional reduction in sales and distribution costs for US airlines from 2000 to 2007 represented another $3 billion of cost savings per year. Unfortunately for the airline industry, what amounts to $13 billion in annual savings did not improve their bottom line by anywhere near the same amount, rather the efficiencies were “competed away” in the form of lower ticket prices. The consumer has the primary beneficiary of these efficiencies, as well as many other efficiencies achieved by the airline industry since deregulation.

35

4.0

Trends in US Airline Productivity

In this section, we describe common measures of productivity used in the airline industry associated with the three most important inputs – aircraft, employees, and fuel. Aggregate measures of total productivity, followed by aircraft productivity, fuel efficiency and employee productivity, are examined for the period since 1978. The impressive gains made by US airlines with respect to all of these measures are further reinforced by the preliminary results of a study of multi-factor productivity (MFP) described in Section 4.6.

4.1

Airline Inputs and Outputs

  In most industries, productivity is typically measured as the amount of output created per unit of input. In the airline industry, outputs are best defined as the capacity created, or Available Seat Miles (ASMs) for passenger operations and Available Ton Miles (ATMs) for cargo operations. On the input side, the most important productive inputs for airlines are capital (aircraft), labor (employees), and fuel. Revenue Passenger Miles (RPMs) is an alternative measure of output, representing the subset of that capacity that is actually consumed. For passenger airlines, ASMs are the best representation of “output” in the strict sense of capacity production. On the other hand, it can be argued that the ability of an airline to fill its ASMs with traffic (RPMs) captures additional facets of efficiency and productivity. In this section, we examine both metrics in the aggregate measures of airline industry productivity, as appropriate.

4.2

Aggregate Measures of Productivity

The  ratio  of  ASM  output  to  total  operating  expenses  provides  an  aggregate  productivity   measure  that  can  be  tracked  over  time,  as  shown  in  Figure  4.1.  Based  on  this  simple  metric,   aggregate  productivity  for  the  US  airline  industry  has  increased  by  over  50%  since  1978.     In  the  early  1980s,  US  airlines  were  producing  just  over  5  ASMs  for  each  dollar  (in  real   2010  terms)  of  operating  expense.    In  the  most  recent  decade,  productivity  has  increased  to   36

nearly  9  ASMs  per  expense  dollar,  a  73%  increase  between  the  least  productive  and  the   most  productive  years,  1981  and  2009,  respectively.  The  increases  in  productivity  are  not   without  fluctuations:  this  aggregate  productivity  metric  reflects  approximately  ten-­‐year   cycles  in  which  overall  productivity  grows  but  year-­‐to-­‐year  productivity  does  not  increase   in  all  years.    The  main  reason  for  this  fluctuation  is  the  volatility  of  fuel  prices  –  ASMs  per   dollar  of  operating  expense  is  lower  for  years  in  which  fuel  prices  surged.    2008  is  the  only   year  in  which  aggregate  productivity  was  at  or  below  the  level  of  the  same  point  in  the   preceding  cycle,  due  to  the  profound  impact  of  historically  high  fuel  prices.Figure  4.1:   Inflation Adjusted Total Operating Expense Productivity   Figure  4.1:  Inflation  Adjusted  Total  Operating  Expense  Productivity 10   9  

ASMs  per  Dollar  Operating  Expense

8  

7   6   5   4   3   2   1  

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

-­‐

     

37

4.3

Aircraft Productivity

A commonly used metric of aircraft productivity is aircraft “utilization,” measured in blockhours per day per aircraft. Block hours refer to the time that an aircraft is in use, beginning when the cabin door is closed and ending when the cabin door is opened at the destination. Block hours therefore include ground taxi times as well as flight times but exclude the “turn-around” times spent by each aircraft at the gate for passenger unloading/loading and aircraft ground servicing. The ability of an airline to achieve a certain level of aircraft utilization depends on the characteristics of its network, its schedule, and its efficiency in turning around an aircraft in between flights. Differences in aircraft turn times can be substantial – Southwest Airlines has made 20-30 minute average turn times a focal point of their low-cost operations (Gittell, 2003), while some NLCs plan for turn times at their connecting hubs as long as 1.5 to 2 hours to allow for passenger connections baggage transfers. Larger aircraft require longer turn times than smaller aircraft, as do aircraft arriving or departing on international services (due to additional customs and security requirements). Figure 4.2 shows aircraft utilization rates for the NLC and LCC groups of US airlines over the past 30 years. LCCs consistently have achieved higher aircraft utilization, by a relatively consistent margin of 1 to 1.5 block hours per day. More point-to-point flights with shorter turnaround times positively affect LCC aircraft utilization rates, while connecting hubs, international services, and even time zone constraints act to limit NLC utilization rates. Overall, aircraft utilization increased dramatically during the period shown, peaking in 2007. NLC aircraft utilization increased by about 10%, while LCC aircraft utilization increased by over 30% through 2007. Both groups were affected by the economic downturn in 2008 and thereafter, as reduced demand and higher fuel prices led to capacity reductions that lowered aircraft utilization. Notable increases in aircraft utilization are evident during three time periods: the initial effects of deregulated route selection and scheduling caused an increase 1982-1987; the economic recovery after the first Gulf War led to consistent increases 1994-2000, and industry restructuring efforts led another period of utilization growth 2004-2007. On the other hand, major decreases in aircraft productivity have been inevitably correlated with economic 38

downturns and high fuel prices. Airlines respond to increased fuel prices and/or declining demand by reducing frequencies and cutting unprofitable routes without being able to remove aircraft from their fleets in the short run, resulting in sometimes dramatic drops in aircraft utilization rates. Figure 4.2: Aircraft Utilization, NLC v. LCC Figure  4.2:  Aircraft  Utilization,  NLC  v.  LCC 14  

Block  Hours  per  Aircraft  Day

12   10   8   6   4   2  

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

-­‐

NLCs

LCCs

While both airline groups saw declines in aircraft utilization beginning in 2001 and again in 2008, the magnitudes and duration of the declines were different. After 2001, LCCs were able to rebound in terms of aircraft utilization much more quickly than the NLCs, returning to pre-2001 utilization levels by 2004. It took NLCs longer to recover, as they were not able to return to their 2000 utilization rates until 2006, which was the first year of US airline industry profitability since 2000. Yet, even after all of the efforts to reduce costs and improve productivity by NLCs during the early 2000s, their aircraft utilization rates remained 10-15% below that of LCCs (although this gap narrowed a little by 2009). Another measure of aircraft productivity is the output (ASMs) generated per aircraft per day. 39

ASMs per aircraft-day is calculated as the product of the number of departures per day per aircraft, the average stage length of these departures, and the number of seats on the aircraft. The evolution of this measure of aircraft productivity is shown in Figure 4.3 as an industry total as well as for NLCs, LCCs, and ECs (Express Carriers). Industry-wide, ASMs per aircraft-day grew immediately following deregulation but have been steadily declining since the mid-1980s, with most recent years reaching levels just below the late 1970s. NLCs have seen slow but steady growth in ASMs per aircraft day, with notable improvement in the mid-2000s, a time in which legacy carriers were determined to improve the efficiency of their operations. LCCs experienced significant growth in ASMs per aircraft-day until the mid-1980s, but since then have shown aircraft productivity levels very near identical to NLCs. Figure 4.3: Aircraft Productivity by Carrier Type for Domestic Operations Figure  4.3:  Aircraft  Productivity  by  Carrier  Type  for  Domestic  Operations 700  

ASMs  per  Aircraft  Day  (Thousands)

600   500   400   300   200   100  

1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

-­‐

Industry  Total

NLCs

LCCs

ECs

It is important to note that the ASMs per aircraft-day productivity metric is heavily affected by the average aircraft size among the airlines in each group. For NLCs with a greater proportion of 40

larger aircraft flying longer routes, we would expect this metric to be higher than LCCs operating smaller narrow-body aircraft on shorter routes. Yet, the similarity of this measure between the two groups suggests that LCCs are able to generate relatively more ASMs with their smaller aircraft. Reducing turn times allows LCCs to offer more departures per day per aircraft, and the lack of first class cabins on most LCCs also increases the number of seats on each departure. The total industry decline in this aircraft productivity metric has been driven by another phenomenon: the growth of the Express Carrier (EC) segment in recent years. ECs typically fly regional jets that are much smaller with fewer seats than the majority of aircraft flown by NLCs and LCCs. Additionally, ECs typically operate short-haul routes as feeders for long-haul routes served primarily by NLCs, resulting in more ground time per aircraft day. The increases in EC aircraft productivity in the last decade stem from a shift away from the smallest aircraft used, namely regional jets with 50 seats or fewer. Despite this recent increase in EC aircraft productivity, the industry average continues to fall because the volume of EC aircraft days has grown so substantially.

4.4

Fuel Efficiency Measures

Accounting for approximately one-third of a typical airline’s cost structure, fuel expenditures are significant and have been growing in recent years. In this section, fuel productivity is examined with two measures of consumption and expense, namely, ASMs per gallon of fuel purchased and ASMs per dollar of fuel expenditures, respectively. As has been mentioned on several occasions, the volatility and unpredictability of fuel prices has been a defining element of the cyclicality of airline industry performance. Figure 4.4 shows the industry-wide average fuel productivity by volume since deregulation, overlaid on the total volume of fuel consumed in the same time period. For reference, RPMs (Revenue Passenger Miles, or the consumed output) is included in this chart as well2. Total fuel consumption has increased steadily over the last three decades, line with the consistent growth in capacity flown, peaking in 2000. Significant declines beginning in 2001 and a subsequent drop                                                                                                                

2  For  the  years  1985-­‐1989,  there  was  a  reporting  inconsistency  in  the  volume  of  fuel  consumed,  resulting  in  a  skewing  

of  the  productivity  scores.  Reasonable  interpolation  provides  confidence  that  also  reporting  practices  might  have   changed  for  that  subset  of  years,  the  remaining  data  is  accurate.  

41

in 2008 reflect capacity cuts, making recent levels of fuel consumption comparable to the levels of the early 1990s.

RPM and ASM fuel productivity has increased dramatically since

deregulation: by 73% for produced output (ASMs) per gallon of fuel and by 128% for consumed output (RPMs) per gallon of fuel. These impressive gains in fuel productivity are the result of a variety of factors, including efforts by airlines to improve capacity management, fleet and schedule planning, as well as load factors. At the same time, new aircraft technologies affecting both engine performance and airframe aerodynamics (e.g., winglets) have contributed to improved fuel efficiency. By 2010, US airlines delivered 64 ASMs per gallon of fuel (and 52 RPMs), meaning the industry’s fuel efficiency exceeds that of automobiles, which on average deliver about 25 miles per gallon and are occupied by fewer than 2 passengers. Figure 4.4: Fuel Consumption Productivity and Total Fuel Consumption Figure  4.4:  Fuel  Consumption  Productivity  and  Total  Fuel  Consumption 70  

20   18  

60  

ASMs/RPMs  per  Gallon

50  

14   12  

40  

10   30  

8   6  

20  

Gallons  of  Fuel  (Billions)

16  

4   10  

2  

-­‐

-­‐ 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Gallons  Consumed

ASMs  per  Fuel  Gallon

RPMs  per  Fuel  Gallon

Series1

Series4

Fuel productivity relative to expense, however, paints a very different – and extremely volatile – picture over the last few decades. Figure 4.5 shows the inflation-adjusted fuel expense productivity, again with both produced (ASMs) and consumed (RPMs) measures overlaid on the 42

total expense. After adjusting for inflation, fuel expenditures were much higher in the early 1980s than many would expect, but they did decline in real terms for the greater part of the following two decades. In recent years, real fuel costs have soared relative to historic levels, peaking in 2008. Figure 4.5: Fuel Expense Productivity and Total Fuel Expense Figure  4.5:  Fuel  Expense  Productivity  and  Total  Fuel  Expense 80  

$60

70  

$50

$40

50   40  

$30

30  

$20

Total  Fuel  Expense  (Billions)

ASMs/RPMs  per  Fuel  Dollar

60  

20   $10

10   -­‐

$0 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Total  Fuel  Cost

ASMs  per  Fuel  Dollar

RPMs  per  Fuel  Dollar

With real fuel expenditures declining or remaining stable during the 1980s and 1990s, while the volume of travel increased substantially, fuel productivity of the US airlines increased significantly. On the other hand, the 2000s have been characterized by a decline in fuel expense productivity such that current levels are comparable to 30 years ago. The relationship between production productivity (ASMs) and consumption productivity (RPMs) has also changed over the years, as load factors have increased. Even with the historically high fuel expenditures (in real terms) during the recent past, both RPMs and ASMs per dollar of fuel expense remain higher than in 1978.

43

4.5

Labor Productivity

Historically, labor related expenses have accounted for as much as 40% of the total operating expenses for US passenger airlines. More recently, this proportion has been reduced, both by airline efforts to reduce labor costs and by the growth of other expense categories such as fuel. In 2010, labor expenses accounted for 29.5% of total US airline operating expenses (excluding transport-related expenses, as discussed earlier), approximately equal to the fuel cost proportion. With increased competition from LCCs, NLCs focused on these expenses in their efforts to cut their unit operating costs. One way in which NLCs have cut total labor expenses is by reducing their total number of employees. Although LCC employment has continued to increase with their continued rapid growth, it has not made up for the dramatic NLC cutbacks. Increased use of outsourcing for functions that have in the past contributed to labor costs is another driver of reduced labor unit costs, resulting in a shift of these costs to the non-labor category. Labor productivity in the airline industry is most typically measured in terms of output per period on a per-employee basis, expressed as the ratio of ASMs to employees. Contrary to theoretical expectations, NLCs with longer average stage lengths and larger aircraft sizes have historically reported lower employee productivity rates than low-cost carriers operating smaller aircraft on shorter stage lengths. Employee productivity of the LCC group of US airlines is about 10% higher than that of NLCs, even as both groups have increased ASMs per employee by more than 35% in recent years (Belobaba, et al. 2009). NLCs achieved this increase in labor productivity through reductions in workforce and relaxation of restrictive work rules. Both NLCs and LCCs have also been able to increase employee productivity by replacing humans with technology – for making reservations, buying tickets, and checking in. As with measures of fuel productivity described above, industry-level labor productivity based on both volume and expenditures were considered in this study but did not result in significantly disparate findings. The evolution of labor force productivity, expressed as ASMs per FTE (full time equivalent employee), is presented in Figure 4.6. The total number of employees in the US airline industry grew with increasing capacity and traffic through the 1980s and 1990s, with declines observed during economic downturns at the start of each decade. However, after peaking at almost 550,000 in 2000, US

44

airline employment plummeted by over 30% by 2010, driven largely by the NLC labor force cuts in the early to mid-2000s. Figure  4.6:  Labor  Force  Productivity  and  Total  Labor  Force  

3.0  

600  

2.5  

500  

2.0  

400  

1.5  

300  

1.0  

200  

0.5  

100  

-­‐

Full  Time  Equivalent  Employees  (Thousand  FTEs)

ASMs/RPMs  per  FTE  (Millions)

Figure  4.6:  Labor  Force  Productivity  and  Total  Labor  Force

-­‐ 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 FTEs

ASMs  per  FTE

RPMs  per  FTE

The productivity of labor has grown in waves since deregulation, and has reached historically high levels. ASMs per FTE have more than doubled, increasing 108% since 1978, with more than half of that gain achieved since 2001 as a result of the restructuring efforts of the largest NLCs. When including the effects of increased load factors, RPMs per FTE have increased even more, by 163%. Overall, labor productivity among US airlines has shown consistent and dramatic improvement since deregulation, and this improvement has been a critical reason for the survival of the airline industry as we know it through the difficulties of the last decade.

45

4.6

Multi-Factor Productivity

  Previous sections have explored industry aggregate measures of operating costs as well as aircraft, fuel and labor productivity, to provide an overview of the many changes experienced by the US airlines since deregulation. This section presents a more detailed approach to measuring productivity, using a Multi-Factor Productivity (MFP) model. Based on this MFP approach, preliminary results of changes in productivity using US airline data from 1980-2010 are presented. This work is part of a forthcoming MIT Master’s thesis (Powell, 2012). Multi-Factor Productivity (MFP) is an approach that can generate aggregate measures of productivity, and has been used in many studies of the airline industry. MFP combines various inputs used in the production process and thus serves as a more comprehensive measure than the single-factor productivity measures described above. In this study, three measures of output were considered: revenue passenger-miles (RPMs), available seat-miles (ASMs) and revenue ton-miles (RTMs) – defined as the amount of tonnage (both passengers and cargo) carried multiplied by the number of miles flown. The inputs required to produce these outputs were categorized as: fuel; capital; labor; and intermediate goods/services. The intermediate category captures components (e.g. interest, insurance, maintenance, etc.) not included in the other categories. Although each of these inputs may be defined using various measures (e.g., gallons of fuel, number of employees) this MFP study defines each input in terms of constant dollars. Previous Studies of Airline MFP There is a large body of literature focused on productivity performance of the airline industry. The studies range from those that used measures of productivity to assess the early effects of deregulation (Caves, Douglas and Tretheway, 1983) to those that compared productivity performance across major airlines worldwide (Oum and Yu, 1996). Baltagi, Griffin and Daniel (1995) looked specifically at the impact of deregulation on operating costs in the airline industry in terms of productivity performance. Homsombat, Xiaowen and Sumalee (2010) used a similar approach to identify key factors that led to changes in productivity and cost competitiveness of North American carriers from 1990-2007. While they found significant productivity 46

improvements over the time period, they also note how efficiency gains have been largely offset by the increase in fuel prices. They also found that productivity improvements are correlated with changes in labor costs. These findings are very much in line with the aggregate productivity changes that we have presented in previous sections. MFP Approach The preliminary results presented in the next section are based on use of the “basic growth accounting” MFP approach to compute aggregate measure of changes in productivity for the US passenger airline industry. The growth-accounting methodology examines the change in output over time and relates it to the combined change in inputs weighted by their reported shares of total inputs. Mathematically, this approach could be best described by the following: ∆! ∆! ΔLabor ΔCapital ΔInt. ΔFuel = −[ ! + ! + ! + ! ] ! ! !"#$% !"#$%"& !"#. !"#$ where: • • • • • •

Δ! ! Δ! !

= Growth of MFP = Growth of gross output

Δ!"#$% !"#

%$= Growth of labor cost

Δ!"#$%"& !"#$%"& Δ!"#. !"#.

= Growth of intermediate cost

Δ!"#$ !"#$

= Growth of capital cost

= Growth of fuel cost

•

! = Share of labor cost in input

•

! = Share of capital cost in input

•

! = Share of intermediate cost in input

•

! = Share of fuel cost in input

The data used comes from MIT’s Airline Data Project which, in turn, was extracted from the Bureau of Transportation Statistics Form 41 database (DOT, 2011). For this study, the relevant 47

data included information on industry capacity, traffic and financial performance from 1978 through the third quarter of 2010. A very small portion of the input data (i.e. other services purchased component) was missing from 1978-1989. Transport related expenses are removed from all calculations, as explained earlier. All inputs are expressed in constant dollars (1980 base). The specific components of each output and input category are discussed in more detail in Apostolides (2008). This fundamental approach gives a gross-output measure of changes in productivity that cannot be attributed to changes in primary cost inputs. In other words, it estimates the changes in total output (ASMs, RPMs, or RTMs) that cannot be attributed to changes in the any of the component inputs (fuel, labor, capital, and intermediate) measured in constant dollars. The resulting increase (or decrease) in productivity is a result of a combination of other factors. Preliminary Results Δ!

In this preliminary study, the time periods used to calculate percent changes in productivity ( ! ) were year-to-year, every 5 years, every 10 years, from 1980 to 2010, and then cumulatively since 1980. The results of the year over year MFP approach for estimating aggregate changes in productivity in the US airline industry are summarized by in Figure 4.7. The graph shows changes in MFP over time using the Growth Accounting methodology outlined above. Because this approach estimates annual changes in productivity, negative productivity growth is shown for the years with unusual events that included fuel crises and economic downturns. The minimal to negative productivity gains in the late 1980’s may best be explained by over-capacity coupled with increasing labor costs and fuel prices. Annual shifts showing significant negative productivity are most typically explained by surges in input costs (typically fuel and/or labor costs) as well as drops in demand leading to reductions in both capacity (ASMs) and traffic (RPMs).

48

Figure 4.7: Year-by-Year Change in Total MFP Since 1980 Figure  4.7:  Year-­‐by-­‐Year   Change   in  Total  MFP  Since  1980 15%

5% 0% -­‐5%

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Year  Over  Year  Percent  Change

10%

-­‐10% -­‐15% -­‐20% -­‐25% RPMs

ASMs

RTMs

The cumulative increases in airline MFP productivity derived from the Growth Accounting methodology for the period since deregulation are shown in Figure 4.8. Irrespective of which measure of “output” is assumed, the US passenger airlines have shown tremendous productivity improvement over the past 30 years. On the basis of ASMs as output, aggregate airline MFP has increased by about 80%. Based on RTMs, cumulative MFP growth is estimated to be close to 140%. Use of the RPM measure increases the estimate of MFP growth since 1980 to a gain of 160%, that is, the aggregate MFP of the US passenger airlines has grown by over 2.5 times when increases in average load factor are included.

49

Figure 4.8: Cumulative Growth of MFP Since 1980 Figure  4.8:  Cumulative  Growth  of  MFP  Since  1980 180%

Percent  Change

140%

100%

60%

-­‐20%

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

20%

RPMs

ASMs

RTMs

The  preliminary  results  of  the  MFP  study  show  that  the  US  airline  industry  has  made  significant  productivity   gains  since  deregulation.  Future  work  will  employ  the  multilateral  estimation  procedures  outlined  in  several  of   the  aforementioned  studies  to  decompose  the  sources  of  changes  in  MFP.    In  addition,  the  study  will  attempt  to   estimate  productivity  differentials  using  regression  models  on  a  number  of  output  and  network  control  variables   such  as  load  factor,  stage  length,  and  fleet  composition.    

50

5.0

Evolution of Airline Networks and Airport Connectivity

Over the past three decades, the route freedoms of deregulation have allowed the networks of US airlines to expand significantly. Network Legacy Carriers (NLCs) have focused their network development on hub-and-spoke operations. New entrant Low Cost Carriers (LCCs) have relied more on point-to-point services that can provide them certain productivity advantages. Yet, as discussed in Section 2.3, LCC and NLC business models are converging. In this section, we summarize the findings of two more detailed MIT studies – one looking at airport connectivity since 1980 (Jenkins, 2011) and the other focusing on the evolution of US airline domestic networks since 2000 (Hernandez et al, 2011). First, the basic differences between hub-and-spoke and point-to-point airline networks are reviewed as background. Then, the findings of Jenkins’ study of changes in measures of passenger accessibility at the top 450 US airports are described. Finally, the impacts on US domestic airline networks of recent developments in the industry are discussed – the extent to which NLCs and LCCs have moved toward increased hub operations, the impacts of recent mergers on domestic networks, and the changing relationships between large carriers’ mainline operations and regional partnerships.

5.1

Hub-and-Spoke versus Point-to-Point Networks

In hub networks, airlines can provide a joint supply of seats to multiple origin-destination (O-D) markets with fewer flight departures and fewer aircraft, with lower total operating costs than in a point-to-point route network. Consider a simple connecting hub network with 20 flights into and 20 flights out of a hub airport, as shown in Figure 5.1. In this hub network each flight leg arriving or departing the hub provides joint supply to 21 O-D markets – one “local” market between the hub and the spoke, plus 20 additional “connecting” markets. The airline can thus serve a total of 440 O-D markets with only 40 flight legs and as few as 20 aircraft flying through the hub. In contrast, a complete “point-to-point” network providing non-stop service to each market could require 440 flight legs and hundreds of aircraft, depending on scheduling requirements (Belobaba et al, 2009). 51

Figure 5.1: Example of Hub Network (Source: Belobaba et al, 2009)

By consolidating demand from many O-D markets on each flight leg into and out of the hub, the airline can provide connecting service even to small O-D markets that cannot otherwise support non-stop flights. This consolidation also allows the hub airline to provide increased frequency of departures from each spoke city, as it likely operates several connecting banks per day in each direction at its hub airport. The large volume of flights at the hub airport can also contribute economies of scale and lower unit operating costs in areas such as aircraft maintenance, crew scheduling, catering, and ground handling. On the other hand, the concentration of airline operations at a large hub can reduce aircraft and crew utilization, compared to point-to-point networks. Hub operations can lead to reduced flexibility in scheduling of departures due to the fixed timing of connecting banks and the need for increased “turn-around” times for aircraft at the hub, to accommodate passenger and baggage connections. And, there can be uneven use of resources (such as gates and runway capacity) at hub airports, with large surges of passenger and flight activity during peak connecting times. There are also potential congestion and delay costs at the hub airport. The fundamental economic advantages of hub networks to airlines – the increased revenues from more frequent (connecting) flight departures combined with the clear unit operating cost savings 52

from operating fewer (and larger) aircraft than in a complete point-to-point network – far exceed their disadvantages. While point-to-point services can be more cost-efficient and, in turn, profitable, in high-density O-D markets that can support non-stop flights, the vast majority of smaller markets cannot be served profitably without the hub-and-spoke network structure that has been the focus of virtually all US airlines since deregulation.

5.2

US Airport Connectivity: A Passenger Perspective Since 1980

This section summarizes the changes in the connectivity of airports as a result of the evolution of US airline networks since 1980, and the increased use of hub-and-spoke operations by the NLCs in particular. We present the major findings of an extensive study of US domestic passenger ticket itineraries, focusing on traffic growth, destinations served, and path quality changes at the top 450 US airports, from the passengers’ perspective. Because the analysis is based on actual passenger tickets, these measures reflect consumers’ demonstrated behavior in selecting airports and choosing itineraries. The details of the data, methodology and a more complete description of findings can be found in Jenkins (2011). Path quality and circuity are two measures of air travel (in)convenience experienced by passengers as they travel from their origin airport to their final destination airport. Average path quality reflects the mix of non-stop versus connecting itineraries flown by passengers at a given set of airports – higher path quality results from more non-stop itineraries, and lower path quality results from more itineraries involving stops and connections3. Recognizing that not all connections are equal in terms of the total travel time required, circuity is another proxy for the inconvenience of a connecting flight. Circuity is calculated as the average distance flown by all passengers in an O-D market, including stops and connections, divided by the actual non-stop distance between the origin and destination. Destinations Flown from each airport is another simpler measure of accessibility. All three metrics are discussed in this section to provide an overview of accessibility changes in the US airport network since 1980. Aggregate measures were compared for various airport classifications in two airport categories, as defined by the                                                                                                                 3  The  distinction  between  itineraries  with  stops  and  connections  lies  in  that  a  one-­‐  or  multiple-­‐stop  itinerary  is  

defined  as  a  journey  that  requires  the  aircraft  to  take  off  and  land  more  than  once  between  the  origin  and  destination,   but  does  not  require  the  passenger  to  transfer  aircraft  at  that  airport  (or  airports).    A  one-­‐  or  multiple-­‐connection   itinerary  requires  a  passenger  to  change  aircraft  at  that  intermediate  airport  (or  airports).  Connections require a minimum 45 minutes extra travel time for passengers and increase the chances of delay, lost luggage, etc.  

53

FAA. The two categories are Commercial Service and Primary Service, and the Primary Service classifications include large hubs, medium hubs, small hubs, and non-hubs. The FAA definitions of these categories and classifications can be found in the Appendix. Path Quality The evolution of aggregate per-passenger path quality for each airport type is shown in Figure 5.2. Based on passenger-weighted path quality, accessibility has increased considerably since 1980 at the smallest Commercial Service airports while decreasing at medium, small and nonhub airports, and remaining roughly constant at large hubs. The smallest airport categories, nonhub and Commercial Service, did not experience path quality declines as significant as the other categories, with Commercial Service airports showing an increase in average path quality. There are two likely reasons for this. First, in the regulated era many smaller airports were served by what amounted to hub-and-spoke service, with regional carriers feeding passengers into larger hub airports served by mainline network carriers. When the deregulated airline industry strengthened its hub-and-spoke structure, passengers at the small airports saw little change. Secondly, the number of destinations available from small airports decreased by over 20% from 1980 to 1990 and by 38% from 1980 to 2010 as airlines eliminated unprofitable routes. Thus, many of the O-D markets with lower path quality were eliminated, elevating the average path quality of the remaining markets. Overall, average path quality has declined since 1980 in US domestic air transportation. However, the decline in this aggregate measure is due in part to the fact that far more passengers are choosing to select a connecting itinerary based on a lower fare. Improved airline efficiency from hubbing has lowered unit operating costs, and increased competition following deregulation has forced the airlines to pass some of that cost savings to consumers in the form of lower fares, as discussed below.

54

Figure 5.2: Passenger Weighted Path Quality

Destinations Flown All airport categories on average have seen decreases in the number of destinations flown by passengers since 1980. In absolute terms, large hubs experienced the largest decrease in destinations, dropping on average by 97 destinations or 20% from 1980-2010. This large absolute decrease is due in part to the fact that they had the greatest number of destinations at the start of the sample period. Commercial service airports experienced the largest average decrease per airport on a percentage basis: 50 destinations (38%) from 1980-2010. Medium, small, and non-hub airports fared somewhat better, decreasing in destinations flown by 15%, 10%, and 4% respectively. Figure 5.9 details the changes in destinations available for each airport category. Again, the decline in destinations served is a direct result of the free entry and exit in individual markets made possible by deregulation.

55

Figure 5.3: Destinations Flown

Circuity The most consistent change seen during the last 30 years is the increase in average circuity of travel in all airport categories between 1980 and 1990, as shown in Figure 5.4. This resulted from the airlines’ significant shift towards the use of hub-and-spoke networks to consolidate traffic in the first decade after deregulation, combined with passengers’ ability to choose connecting itineraries that offered lower fares.

56

Figure 5.4: Passenger Weighted Circuity

Following 1990, the changes are less consistent across airport types with an overall percentage increase in circuity from 1980-2010 between 0 and 5%. The average circuity of itineraries from an airport increases as the size of the airport decreases, for two principal reasons. First, smaller airports rely more on larger hub airports for connecting flights because local market demand cannot support non-stop flights. Second, market distance (the non-stop distance from origin to destination) is shorter at smaller airports and shorter trips tend to have greater circuity. By multiplying the passenger-weighted circuity by the average passenger-weighted, non-stop market distance we can estimate how many excess miles were flown by the average random passenger, as shown in Table 5.1.

57

Table 5.1: Average Excess Miles Flown 1980

1990

2000

2010

Large hub

28

36

39

40

Medium hub

34

44

49

64

Small hub

57

72

92

100

Non-hub

39

75

107

108

Commercial Service

56

75

82

98

In 2010, at all airport types, passengers flew further in excess of non-stop distances than they had in 1980, from 12 more miles per passenger at large hubs to 69 more miles at non-hubs. The changes in circuity distances for large and medium hubs are quite small, assuming jet aircraft are serving many of these flights. At smaller airports, particularly those served by turbo-prop aircraft, an increase in circuity of about 50 excess miles translates into about 10 minutes of increased flying time at 300 mph turbo prop speeds, which is a considerable amount of time when summed across all passengers. The circuity presented in this analysis is based on all O-D markets whether they were served with non-stop flights or connecting itineraries. If circuity were calculated based only on connecting flight itineraries, these values would most certainly be higher. Average Fares The growth of hub-and-spoke networks has not only allowed airlines to operate more efficiently, it has increased the effective competition in many domestic O-D markets. With increasing market presence by LCCs, competition increased further leading to lower average fares. Across all airport categories, fares have declined by dramatic amounts since 1980, even when adjusting for inflation. In 1980 average one-way ticket prices were between $270 and $323 (inflation adjusted to 2010) depending on airport size. Through 1990 and 2000, average fares declined steadily and by 2010 they averaged between $151-$184, a drop of 41-45%. Figure 5.5 displays the steady and dramatic decline of fares across all airport types. 58

Figure 5.5: Inflation Adjusted Average Passenger Fares

Fares have decreased even more on a per-mile basis because passengers are on average flying to more distant destinations than they were in 1980. The average market distance increased by 19% at large hubs, 35% at medium and small hubs, 44% at non-hubs, and 19% at Commercial Service airports. In all airport categories except for Commercial Service, itineraries less than 500 miles decreased on a percentage basis while trips of 500 or more miles increased. At large hubs the percentage of sub-500-mile itineraries sank from 30% in 1980 to 17.4% in 2010, at medium hubs they were down to 27.7% from 48.7%, small hubs fell from 44.7% to 23%, and non-hubs dropped to 19.5% from 43.6%.

59

5.3

US Airline Domestic Network Development Since 2000

Given the growth of hub networks used by NLCs and the emergence of new entrant LCCs that have been assumed to focus more on point-to-point operations, we undertook a more detailed study of US domestic flights since 2000. Operations data for US domestic airline flights operated in May 2000, 2005 and 2010 were analyzed to explore the changes in US airlines’ networks, their dependence on the hub-and-spoke model, shifts to regional partners for domestic flying, and the impacts of recent mergers of large airlines. We summarize the major findings of this study in this section, while the details of the methodology, data and analysis can be found in Hernandez et al (2011). Over the last decade, the US airline industry has seen significant consolidation, with the majority of domestic flights being operated by one of the three leading carriers: Southwest, Delta and American. Figure 5.6 displays the evolution of leading carriers by mainline domestic flight volumes since 2000. Figure 5.6: Mainline Domestic Flight Volumes by Carrier

Monthly  Domestic  Mainline  Flights

Figure  5.6:  Mainline   Domestic  Flight  Volumes  by  Carrier 100,000   80,000   60,000   40,000   20,000   0  

2000

2005

2010

60

The dependence of US airlines on the hub-and-spoke network model has grown since 2000, with 80% to 100% of domestic flights arriving or departing a designated hub for most major US carriers. Even Southwest Airlines, assumed by many to be a point-to-point carrier, now operates a network much more focused on network connections, with over 50% of its flights arriving or departing a designated hub. Table 5.2 contains as a weighted average (by flight volume) the percentage of hub flights for major US carriers. With Southwest’s relatively low hub dependence included, 83% of all flight segments either arrived at, departed from, or both arrived at and departed from one of that carrier’s designated hubs in a sample month during 2010. LCCs have demonstrated similarly high hub dependence throughout the decade, even when compared to NLCs, presented also in Table 5.2. LCC hub utilization of 90% contradicts the theory that these carriers offer primarily point-to-point service. Despite the fact that many LCCs initially choose to serve point-to-point markets, organic growth proves to be difficult without resorting to a hub network structure. Table 5.2: US Airline Domestic Hub Operations: LCC v. NLC (Weighted Average) 2000

2005

2010

Southwest

40%

46%

52%

LCCs, excluding Southwest

95%

90%

90%

NLCs

75%

92%

97%

Total US Airlines

80%

84%

83%

As shown in Figure 5.7, most US carriers experienced very high levels of hub use throughout the decade, with even the most hub-dependent airlines increasing their percentage of hub flights between 2000 and 2010.

61

Figure 5.7: Domestic Mainline Percentage by Carrier Figure  5.7:  Domestic  Mainline   Percentage  by  Carrier

Percent  Hub  Flights

100% 80% 60% 40% 20% 0%

2000

2005

2010

Eight of the top carriers shown experienced an increase in hub dependence throughout the decade, while Continental and Frontier declined slightly but remained above 95% in 2010. JetBlue, AirTran and Frontier showed a slight decrease in hub flights over the decade, but these LCCs still operated well over 80% of their domestic flights in May 2010 into or out of a designated hub. Frontier, AirTran, and JetBlue all experienced significant growth4 during the same time period, increasing flight volumes by a factor of 2.3, 2.5, and 30.6, respectively. The only other carrier to post an increase in flight volumes during the same period was Southwest, with a 27% increase; the airline’s hub dependence also grew from 40% to 52%. Regional Partnerships The relationship between US mainline carriers and regional service providers varies and creates difficulty in identifying partnerships from the available operational data. Some NLCs have wholly owned subsidiary airlines that provide regional feeder services, making it easier to                                                                                                                 4  Frontier’s  growth  was  largely  a  function  of  its  2009  acquisition  by  Republic  Airways  after  its  2008  Chapter  11  

bankruptcy  filing.    AirTran  and  JetBlue’s  growth  was  primarily  organic,  particularly  in  the  case  of  JetBlue  since  its   founding  in  1999  resulted  in  very  low  flight  volumes  in  early  2000  and  consequently  a  higher  percent  change  in   subsequent  years.  

62

associate regional flight volumes to the appropriate NLC. There exist other regional airlines that provide regional services to multiple – and often competing – NLCs. These independent carriers may have exclusive agreements with a particular NLC at each airport it serves; in other cases, the same regional carrier might provide service for multiple, competing NLCs to and from the same airport, even on the same flight segment. Two selected NLC/subsidiary pairs, American/American Eagle and Delta/Comair, saw major declines in total domestic flight volumes throughout the decade. American’s declines came primarily from a reduction in its mainline operations, while Delta’s drop off stemmed mostly from fewer Comair operations. Different operating strategies by these NLCs were reflected in the use of regional partners at connecting hubs: Comair plays a nearly nonexistent role at Delta primary hub ATL, while AE’s role has increased at primary American hubs DFW and ORD. On the other hand, Comair’s role grew at Delta secondary hubs, and AE’s presence shrunk at American secondary hubs. At most key hub airports, hub carriers maintained or increased their share of total flights over the decade with the increased use of regional partners for domestic feeder flights. The majority of many of the regional carriers’ flights are marketed by the larger hub carriers, and although mainline operations have declined, the total representation of those hub carriers has remained high. As one example, Philadelphia International Airport (PHL) is served by an increasing number of carriers with an increasing role, as shown in Figure 5.8. Figure 5.9 then shows the effective market share of the main hub carrier, US Airways, based on regional partnerships. Although the volume of US Airways domestic mainline flights at PHL has decreased, the overall share of the US Airways brand has increased with more domestic flights operated by regional partners.

63

Figure 5.8: PHL Key Carriers Figure  5.8:  PHL  Key  Carriers

Monthly  Domestic  Flights

40,000  

30,000  

Other Mesa  Airlines Air  Wisconsin  Airlines

20,000  

PSA  Airlines Piedmont  Airlines Republic  Airlines Southwest  Airlines

10,000  

US  Airways

0   2000

2005

2010

Figure 5.9: PHL Hub Carriers Figure  5.9:  PHL  Hub  Carriers 30,000  

Monthly  Domestic  Flights

25,000   20,000   15,000   10,000   5,000   0   2000

2005

US  Airways*  [Republic,  Piedmont,  AirWisc,  Mesa,  PSA]

2010 Other

64

A similar shift of flight volumes from mainline carriers to regional partners was observed at nearly all major NLC hub airports since 2000, accompanying the increasing market share of hub carriers relative to all other service providers at those same airports. Exceptions to these findings include Los Angeles International Airport (LAX) and LaGuardia Airport (LGA). Also, Pittsburgh International Airport (PIT), Cincinnati/Northern Kentucky International Airport (CVG), and Lambert-St. Louis International Airport (STL) have been abandoned as designated hubs due to market conditions, strategic initiatives, and mergers. Impacts of Mergers In the wake of intense fare competition, bankruptcies, and fuel crises, a wave of mergers between major US carriers occurred in the last decade in the pursuit of cost efficiency through greater scale economies. Key mergers have had significant impacts on flight volumes, regional partner relationships, and hub utilization. Figures 5.10 and 5.11 show the impact of recent mergers on mainline domestic flight volumes for two of the largest mergers completed since 2000. US Airways and Delta Airlines effectively absorbed the domestic mainline flight volumes of America West and Northwest Airlines, respectively. This resulted from not only the consolidation of operations, but also the shift from mainline NLC flights to NLC-branded, but regional carrier-provided flights so important toward the end of the decade. For both US Airways/America West and Delta/Northwest, the acquiring airline maintained its hubs post-merger, and both transitions were accompanied by a significant increase in hub dependence. In 2000, only 62% of US Airways’ total flights included one or more designated hubs; this increased to 90% in 2005 and ultimately to 97% in 2010. Delta underwent a similar effort to refocus on hubs: beginning with only 79% in 2000, hub dependence increased to 92% in 2005 and 96% by 2010. Hub operations can be expected to remain high and potentially even continue to increase as a result of industry consolidation.

65

Figure 5.10: US Airways, America West Domestic Mainland Flight Volumes Figure  5.10:  US  Airways,  America  West  Domestic  Mainline  Flight  Volumes

Monthly  Domestic  Mainline  Flights

80,000  

60,836  

60,000  

37,540  

40,000  

20,000  

18,119  

33,970  

16,860  

0   2000

2005 America  West

2010

US  Airways

Figure 5.11: Delta, Northwest Domestic Mainline Flight Volumes Figure  5.11:  Delta,  Northwest  Domestic  Mainline  Flight  Volumes

Monthly  Domestic  Mainline  Flights

80,000  

75,097  

58,750  

60,000   44,657  

61,559  

42,414  

40,000  

20,000  

0   2000

2005 Northwest  Airlines

2010

Delta  Airlines

In the case of the more recent United/Continental merger, United mainline flight volumes have declined over the decade, as shown in Figure 5.12, making it much more difficult for United to 66

completely absorb a similar volume5 of Continental flights. The Southwest/AirTran merger is also unlikely to follow the same pattern of mainline flights being absorbed for various reasons. The AirTran acquisition enables Southwest to compete at Hartsfield-Jackson Atlanta International Airport (ATL), a previously untapped market for Southwest; both Southwest and AirTran flight volumes continue to rise6; and neither Southwest nor AirTran relies on regional partners to provide feed to the same extent as NLCs like US Airways and Delta Airlines. Figure 5.12: United, Continental Domestic Mainline Flight Volumes Figure  5.12:  United,  Continental  Domestic  Mainline  Flight  Volumes

Monthly  Domestic  Mainline  Flights

80,000   64,321   60,000   42,713   40,000  

32,403  

28,750  

24,710   19,900  

20,000  

0   2000

2005 Continental  Airlines

2010

United  Airlines

 

                                                                                                                5  In  May  2010,  United  operated  28,750  mainline  domestic  flights  to/from  ASPM  airports;  Continental  operated  

19,900  during  the  same  period.  

6  The  Wall  Street  Journal  recently  published  an  article  claiming  over  10%  growth  in  traffic  in  May  2011  over  May  

2010,  the  most  recent  data  included  in  this  study.  Traffic  growth  was  partially  due  to  capacity  increases  and  partially   due  to  increased  load  factors.  http://online.wsj.com/article/BT-­‐CO-­‐20110607-­‐707159.html?mod=dist_smartbrief  

67

5.4

Summary

The US airline industry’s dependence on the hub-and-spoke network model has continued to increase, reaching an unprecedented 96% industry average hub use for domestic flights (83% when including Southwest). Even low cost carriers utilize a designated hub for over 90% of all flight segments, excluding Southwest Airlines. Southwest continues to display lower levels of hub dependence, but has grown significantly in the last decade, with over 50% of flights in May 2010 arriving or departing a designated hub. Hub operations have increased for many individual carriers, through mergers, and as a result of high-impact relationships with regional carriers. The effect on passengers has been lower path quality and higher circuity – both suggesting increased inconvenience of traveling for nearly all airport classifications. However, increased inconvenience has provided operational advantages to carriers; when coupled with growing competition, passengers have enjoyed considerably lower fares as a result of increased hub-andspoke dependence.

68

6.0

Looking Ahead: US Airline Industry Challenges

The US passenger airline industry has undergone tremendous change since deregulation, with many of the most important changes to business practices, cost efficiency and productivity summarized in this report. Many stakeholders have been affected in different ways by the industry’s transformation, but the one clear winner has been the air travel consumer. While consumers have benefited from increased competition, lower fares, new entry and innovative service options, airlines have not been able to retain the financial benefits from the many cost and productivity efficiencies they achieved. Perhaps the clearest losers in this transition have been the shareholders of airline companies. Although the US airline industry was deregulated more than 30 years ago, it remains subject to substantial government intervention and regulation. Because the industry is so visible and important to the economy, legislative and regulatory bodies insist on being involved in airline commercial decisions, generally in the name of protecting competition and the consumer. Airline labor relations are governed by the Railway Labor Act, much of which has remained unchanged since 1934. Infrastructure limitations pose an impediment to operational efficiency, in many cases due to regulatory requirements and the inability of major stakeholders to resolve policy differences. Despite all of the efforts of US carriers to restructure themselves in recent years, the remnants of 60 years of regulation continue to affect airline business practices and operations. Looking ahead, it is important to understand the extent to which global rather than domestic competition will shape the future airline industry. The US airline industry has historically been the largest in the world and has been a leader in terms of technology, competition and innovative business practices. This leadership position has been eroded, particularly during the past decade. While the US industry stagnated as it focused on the restructuring of costs and productivity, airlines in other regions of the world have continued to grow. Economic growth and increasing air travel demand sparked aggressive expansion by airlines in Latin America, China, and India, while most carriers in Europe and Asia continued to grow more slowly. At the same time, several rapidly growing airlines based in the Middle East have become important competitors in

69

international O-D markets connected by their well-positioned hubs. Many of these non-US airlines have been able to remain profitable while US airlines were struggling. In the US industry, the NLC and LCC operating and business models have been converging, and with recent consolidation the two types of carrier will continue to co-exist. Some LCCs will begin to expand their international services, perhaps even to transatlantic markets, providing a new competitive challenge to both US and non-US legacy carriers. However, LCCs will not be able to replace the extensive networks and services provided by large NLCs. Constraints on infrastructure capacity – both airports and airspace – and the costs of maintaining and expanding this infrastructure are critical problems for the future of the airline industry. While the FAA has been working toward increasing the capacity of the en route airspace, increasing flight delays suggest that the US air traffic infrastructure is not keeping pace with the demand for air travel. Without major investments in the new technologies of NextGen as well as additional airport infrastructure, it will be difficult to accommodate the expected growth in air traffic, raising concerns for nearly every stakeholder group in the airline industry. Similarly, concerns about the environmental impacts of aviation are growing, and will need to be addressed. Continued movement by governments toward more stringent environmental regulations will encourage the development of new aircraft technologies and force airlines to adopt them more quickly. However, without a return to sustained profitability, it will be difficult for many carriers to fund the investment required to renew their fleets. The US airline industry is highly leveraged. Capital was plentiful and readily available to the industry until the past decade, although the capital that seemed to be always available was not always smart capital. Rather, capital was made available to protect prior investments put at risk by an event or series of events undermining an airline’s performance. This practice appears to be changing as US airlines become more focused on de-leveraging and repairing balance sheets. Historically, the US airline industry has not even been able to earn its weighted average cost of capital. Today, many airlines have decided that investments must at least earn the cost of capital, requiring financial discipline that the industry has not previously exhibited.

70

Our study of productivity has highlighted that the airline business is both capital-intensive and labor-intensive, and is subject to a tremendous cyclicality driven primarily by economic forces and volatile fuel prices. As US airlines become increasingly global in nature, geopolitical events will have as great an impact on the industry as economic conditions, contributing further to this volatility. Repeated cycles of record profitability followed by huge losses have left many US airlines in a weakened financial situation. Given that many elements of this cyclicality are not likely to change, perhaps the greatest challenge to the airline industry is to achieve sustained profitability and greater stability.

71

REFERENCES

Air Transport Association of America (2010), Economic Report 2010, www.airlines.org. Apostolides, A. (2008), "A Primer on Multifactor Productivity: Description, Benefits, and Uses”. US Department of Transportation, Research and Innovative Technology Administration, Bureau of Transportation Statistics, Office of Advanced Studies, Washington, DC. Baltagi, B., Griffin. J., and Daniel, R. (1995), "Airline Deregulation:The Cost Pieces of the Puzzle," International Economic Review 36.1, 245-258. Belobaba, P., Odoni, A., and Barnhart, C. (editors) (2009), The Global Airline Industry, John Wiley and Sons. Calio, N.E. (2011), “Proposed Airline Taxes Would Harm Economy, Jobs”, www.rollcall.com, July 28, 2011. Caves D., Christensen, L., and Tretheway, M. "Productivity Performance of US Trunk and Local Service Airlines in the Era of Deregulation," Economic Inquiry 21, 312-324. General Accounting Office (1993), “Airline Competition: Higher Fares and Less Competition Continue at Concentrated Airports”, Report to Committee on Commerce, Science and Transportation, US Senate, Report GAO/RCED-93-171, Washington, DC, July. Gittell, J.H. (2003), The Southwest Airlines Way, McGraw-Hill, New York. Hernandez, K., Swelbar, W., and Belobaba, P. (2011), “US Airline Domestic Network Evolution 2000-2010”, Working Paper, International Center for Air Transportation, Department of Aeronautics and Astronautics, MIT. Homsombat, W., Xiaowen, F., and Sumalee, A. (2010), "Policy Implications of Airline Performance Indicators: Analysis of Major North American Airlines," Journal of the Transportation Research Board 2177, 41-48. Jenkins, J. (2011), “The Evolution of Passenger Accessibility in the US Airline Industry 19802010.” MST thesis, Massachusetts Institute of Technology, Cambridge, MA. Karlsson, J. (2010), “Airline Ticket Tax Project”, MIT Global Airline Industry Program, http://web.mit.edu/TicketTax/ Oum, T. H. and Yu, C. (1995), "A Productivity Comparison of the World's Major Airlines", Journal of Air Transport Management 2.3-4, 181-195. Powell, R. (forthcoming, 2012), “Measuring Productivity of US Airlines”, MST Thesis, Massachusetts Institute of Technology, Cambridge, MA.

72

Tsoukalas, G., Belobaba, P. and Swelbar, W. (2008), “Cost Convergence in the US Airline Industry: An Analysis of Unit Costs 1995-2006”, Journal of Air Transport Management 14, 179187. United States Department of Transportation (2011), “Air Carrier Financial Reports (Form 41 Financial Data)”, Bureau of Transportation Statistics, Washington, DC. http://www.transtats.bts.gov/

73

APPENDIX A.1

Hub-and-Spoke Analysis Notes (Hernández et al, 2011)

The following carriers were included in the hub-and-spoke analysis, under the following classifications: •

NLCs: Alaska, American, America West, Continental, Delta, Northwest, United, US Airways

•

LCCs: AirTran, Frontier, JetBlue, Southwest

The designated hubs for each of the included carriers are as follows: CARRIER

DESIGNATED HUBS

AirTran Airways

ATL, MCO, BWI, MKE

Alaska Airlines

SEA, PDX, ANC

American Airlines (and American Eagle)

DFW, ORD, MIA, LAX, JFK

Continental Airlines

IAH, EWR, CLE

Delta Airlines (and Comair)

ATL, MSP, DTW, SLC, JFK, MEM, CVG, LAX

Frontier Airlines

DEN

JetBlue Airways

JFK, BOS, MCO, FLL

Northwest Airlines

MSP, DTW, MEM

Southwest Airlines

LAS, MDW, PHX, BWI, DEN

United Airlines

ORD, DEN, SFO, IAD, LAX

US Airways

CLT, PHX, PHL, DCA, PIT

Partnerships between mainline carriers and regional carriers for each airport included in the study are as follows: • •

ATL CLT

Continental: Colgan Air, Expressjet US Airways: Piedmont, PSA, Mesa, Republic, CCAir (Sunbird), Air Wisconsin 74

•

CVG

• •

DEN DTW

• •

EWR IAD

• • •

IAH LGA MSP

• • •

ORD PHL PIT

•

STL

A.2

Delta: Chautauqua, Skywest, Mesaba, Pinnacle, Freedom, Atlantic Southeast, Expressjet, Mesa United: Mesa, Skywest, Great Lakes, Air Wisconsin; Frontier: Republic Delta: Freedom, Mesaba (2010), Pinnacle (2010); Northwest: Mesaba (2000, 2005), Pinnacle (2000, 2005) Continental: Expressjet, Chautauqua, Colgan Air, Commutair United: Shuttle America, Colgan Air, Expressjet, Air Wisconsin, Mesa, Atlantic Coast Continental: Colgan Air, Expressjet US Airways: Allegheny, Piedmont Delta: Compass, Skywest, Mesaba (2010), Pinnacle (2010); Northwest: Mesaba (2000, 2005), Pinnacle (2000, 2005) United: Air Wisconsin, Expressjet, Mesa, Skywest, Chautauqua US Airways: Republic, Piedmont, Air Wisconsin, Mesa, PSA US Airways: Piedmont, Trans States, PSA, Colgan Air, Chautauqua, Air Midwest, Allegheny American: RegionsAir (2005, 2010), Chautauqua (2005, 2010); TransWorld: RegionsAir (2000), Chautauqua (2000)

Airport Connectivity Study Notes (Jenkins, 2011)

The FAA airport classifications included in this study are as follows: • Commercial Service: publicly owned airports that have at least 2,500 but no more than 10,000 passenger boardings each year and receive scheduled passenger service. • Primary Service: are commercial service airports that have more than 10,000 passenger boardings each year. Within the primary service category the FAA further classifies airports into hub size7: o Large hub: 1% or more of annual US passenger boardings o Medium hub: At least 0.25% but less than 1% of annual US passenger boardings o Small hub: At least 0.05% but less than 0.25% of annual US passenger boardings o Non-hub: Fewer than 0.05% of annual US passenger boardings

                                                                                                                7  There  are  29  large  hub  airports,  37  medium  hub  airports,  71  small  hub  airports,  and  239  non-­‐hub  airports  as   classified  by  the  FAA.   75

NOTE: General aviation airports are excluded from the study8.

                                                                                                               

8  In  March  2011,  the  FAA  recognized  375  primary  service  airports,  109  commercial  service  airports,  and  over  2,000  

general  aviation  airports.  

76