The operational impacts of governmental restructuring of the airline industry in China

Transcription

1 Fairfied University Business Facuty Pubications Chares F. Doan Schoo of Business The operationa impacts of governmenta restructuring of the airine industry in China Car A. Scheraga Fairfied University, Copyright 2006 Transportation Research Forum A rights reserved. Peer Reviewed Repository Citation Scheraga, Car A., "The operationa impacts of governmenta restructuring of the airine industry in China" (2006). Business Facuty Pubications Pubished Citation Scheraga, Car A. "The operationa impacts of governmenta restructuring of the airine industry in China." In Journa of the Transportation Research Forum, vo. 45, no. 1, pp This Artice is brought to you for free and open access by the Chares F. Doan Schoo of Business at DigitaCommons@Fairfied. It has been accepted for incusion in Business Facuty Pubications by an authorized administrator of DigitaCommons@Fairfied. For more information, pease contact digitacommons@fairfied.edu.

2 Transportation Research Forum

3 The Operationa Impacts of Governmenta Restructuring of the Airine Industry in China In Juy 2000, the Civi Aviation Administration of China (CAAC) caed for the consoidation of the 10 state-owned air carriers into three groups, headed by Air China, China Eastern, and China Southern. A few months ater in November 2000, the State Counci of China mandated that this consoidation be accompished by the third quarter of As part of this mandate, the CAAC yieded its management contro of air carriers with its focus now being on safety and reguatory issues. Furthermore, the CAAC was required to divest itsef of assets hed in many of the state-owned airines and its interests in more than 120 airports around China, except Beijing Capita Airport. Utiizing data from the Internationa Civi Aviation Organization for 2003 and 2004, this study investigates the operationa impacts of this industry restructuring. The reative operationa efficiency of Air China, China Eastern, and China Southern is compared to a sampe of Asian, European and United States fag carriers. Data enveopment anaysis is utiized to derive efficiency scores for individua airines. The operationa efficiency mode used in this study is derived from that utiized by Schefczyk (1993). The underying structura drivers of efficiency are then investigated via a tobit anaysis with impications for manageria poicy discussed. by Car A. Scheraga INTRODUCTION Before 2001, the Chinese commercia aviation industry was overseen and controed by the Civi Aviation Administration of China (CAAC). The Chinese government impemented a comprehensive set of reguations and poicies that covered a aspects of airine operations. This reguatory regime approved domestic, regiona, and internationa route aocations. Contros were put into pace that set guideines for pubished fares and aircraft acquisition, as we as standards for aircraft maintenance, jet fue prices, airport operations, and air traffic contro. In addition, the CAAC owned some of the country s argest airines. As detaied by Efendiogu and Murray (2003), by 2001 China had 31 internationa, regiona and domestic carriers. These incuded 10 CAAC airines and 21 Provincia airines. The CAAC airines controed about an 80.5% share of the passenger market and an 84.7% share of the cargo market. Furthermore, the CAAC had an average 85% ownership share in its airines. The priority of airines was to buid market share with profitabiity often compromised. In addition to being impacted by operationa inefficiencies, profits were aso significanty affected by increase in fue costs, ow oad factors, and the inabiity of many Chinese airines to structure routes that aowed for the achievement of economies of scae. An exacerbating factor was China s attempt to become part of the Word Trade Organization (WTO), which increased the opportunities for foreign carriers to enter the Chinese market. In Juy 2000, the CAAC caed for the merging of the 10 arge state-owned airines into three groups to be headed by Air China, China Southern Airines, and China Eastern Airines. This was formaized in November 2000 when the State Counci of China issued directive for the merger of these airines in a forma restructuring of the Chinese airine industry. Part of this directive decreed that the CAACP woud yied management contro of airines, instead focusing on safety and reguatory issues. Air China consoidated with China Nationa Aviation Corporation, China Southwest Airines, and Zhejiang Airines. China Eastern Airines absorbed Air Great Wa, China 71

4 Airine Industry in China Yunnan Airines and China Northwest Airines. Simiary, China Southern Airines absorbed China Northern Airines, China Xinjiang Airines and Zhongyuan Airines (Efendiogu and Murray 2003). This study benchmarks the operationa efficiency of the consoidated Chinese airines against a set of goba fag carriers with significant internationa operations for the This set is detaied in Tabe 1. The year 2003 is the most recent one avaiabe in terms of the data utiized. In addition, it is the first year for which consoidated operations data was avaiabe. Thus the study is preiminary in nature, representing a starting point for the deveopment of operations strategies for the three consoidated Chinese airines. A significant factor which must be kept in mind throughout the discussion of this study is the SARS virus epidemic that severey impacted Chinese airines. Passenger numbers dropped, fights canceed, and deiveries of new aircraft were deayed. However, it shoud aso be noted that other airines in the region aso fet the impact of the vira outbreak. Major carriers affected incuded Japan Airines and Thai Airways. MODEL FRAMEWORK AND VARIABLES The behaviora mode utiized to describe airine operationa efficiency is that empoyed by Schefczyk (1993). He defines a framework described by two outputs and three inputs. The outputs are (1) revenue passenger-kiometers and (2) Non-passenger revenue ton-kiometers. The inputs are (1) avaiabe ton-kiometers, (2) operating cost and (3) non-fight assets. Avaiabe ton-kiometers refect avaiabe aircraft capacity. The figure for operating cost refects operating cost excuding capita and aircraft cost captured by avaiabe tonkiometers. Non-fight assets refect a assets not incuded in avaiabe ton-kiometers such as faciities, reservation systems and current assets. Revenue passenger-kiometers are the sum of revenue passenger-kiometers for schedued and charter (non-schedued) services. Nonpassenger revenue ton-kiometers incude 72 Tabe 1: Sampe Airines NAME ICAO ID Aeroines Argentinas ARG Aeromexico AMX Air Canada ACA Air China CCA Air France AFR Air India AIC A Nippon Airways ANA American Airines AAL Asiana Airines AAR Austrian Airine Group AUA British Airways BAW Cathay Pacific CPA China Eastern Airines CES China Southern Airines CSN Continenta Airines COA CSA Czech Airines CSA Deta Airines DAL EL AL ELY Iberia IBE KLM KLM Korean Air KAL LOT Poish Airines LOT Lufthansa DLH Maaysia Airines MAS Mexicana MXA Northwest Airines NWA Pakistan Internationa Airines PIA Scandinavian Airines SAS Singapore Airines SIA SriLankian Airines ALK TAP Air Portuga TAP Tarom ROT Thai Airways Internationa THA United Airines UAL ton-kiometers for freight and mai for both schedued and non-schedued services. Avaiabe ton-kiometers are a sum of avaiabe ton-kiometers for schedued and charter (nonschedued) services. Operating cost is computed as tota operating expenses minus aircraft rent, depreciation, and amortization. Non-fight assets are computed as tota assets minus fight equipment at cost, purchase deposits for fight equipment, and fight equipment under capita

5 eases at cost, with accumuated depreciation for fight equipment and accumuated depreciation for fight equipment under capita eases added back in. Data enveopment anaysis (DEA), discussed beow, is used to assess the reative efficiency of individua airines described by this behaviora mode. This study investigates the impact of a set of operationa and environmenta variabes which previous studies have shown to affect operationa efficiency (Caves et a. 1984; Banker and Johnston, 1993; Schefczyk 1993; Siau and Van Lindt 1997; and Fethi et a. 2002). These incude: average fight ength, passenger revenues as a percentage of tota revenues, schedued service revenues as a percentage of tota revenues, internationa passenger revenue kiometers as a percentage of tota passenger revenue-kiometers, and average oad factor. These variabes, in effect, describe the environment in which an airine operates. Average fight ength captures economies of distance which posits that there is a negative correation between average fight ength and unit cost. For a given aircraft size, increasing the distance of a fight resuts in arger output voume as measured either in passenger revenue-kiometers or ton-kiometers. It must be noted that empiricay this suggested effect has been shown to be ambiguous (Caves et a and Tretheway 1984). The passenger focus for an airine is described by passenger revenues as a percentage of tota revenues. As Oum and Yu (1999) note, air cargo accounts for a arge portion of tota output for many Asian and European carriers based in export-oriented countries. U.S. carriers have traditionay been primariy passengerfocused in their operations. Cargo service is seen as requiring ess input than passenger services, but it generates ess revenue. Specific operationa advantages in carrying cargo as opposed to passengers have been suggested by O Connor (2001). Cargo is usuay carried oneway whie passengers usuay trave roundtrip. Passengers have a preference for day trave whie cargo generay moves at night. Unike passengers, shippers and recipients of cargo are not concerned about indirect routes or pane changes as ong as the cargo arrives when Airine Industry in China expected. Additionay, the aesthetics of aircraft environment are not a concern in the transport of cargo. Schedued service revenues as a percentage of tota revenues are anticipated to have a positive impact on operationa efficiency. Schedued fights require different product and marketing faciities than unschedued charter fights. An increase in the percentage of reguary schedued services aows for a rationaization of operationa routines eading to greater overa efficiency if the necessary concomitant resources and systems are in pace. The internationa focus of an airine is captured by internationa passenger revenue kiometers, as a percentage of tota passenger revenue-kiometers. A priori, what the impact of this measure shoud be on operationa efficiency is not unambiguous. On the one hand, Fethi et a. (2002) suggest that increase in the internationa focus of an airine exposes it to spatia disparities in its operating environment. In structuring biatera agreements, the internationa air transport system has tended to focus on individua or sma sets of routes between countries. This has impeded the achievement of high eves of efficiency over goba networks of air services. There are unresoved issues with regard to ownership and contro, cabotage and the right of estabishment. There is sti divergence across geographic regions with regard to competition aw and poicy in air transport. There are differences in fisca poicies with air transport being subjected to many taxes which finance genera governmenta expenditure. Customs cearance can impede both speed and reiabiity. Finay, airport infrastructure constraints can significanty affect the eve of competition in particuar markets. On the other hand, internationa carriers have been abe to take advantage of strategic aiances with other goba competitors (Dana and Dana 1998). One of the eary exampes of such strategies is the reationship between KLM and Northwest Airines. This aiance began in 1989 when KLM purchased onefifth of Northwest Airines. An open-sky treaty between the Netherands and the United States gave both airines unrestricted rights between 73

6 Airine Industry in China their respective countries. The two airines impemented a joint marketing program, a goba business cass program, and codesharing arrangements. In addition, each of the two airines was aso cooperating cosey with other respective partners. Biatera agreements have proiferated in the industry. In 1998, when reguations were reaxed between Japan and the United States, American Airines and Japan Airines began code-sharing. This agreement gave Japan Airines access to amost 300 airports served by American Airines and its subsidiary American Eage. Simiary, destinations of Japan Airines were opened to American Airines. Japan Airines is particuary interesting in that its fights use aircraft, crew, and duty-free faciities of numerous other arge goba carriers. Strategic aiances have aowed goba airines to coordinate their fights and to cooperate in aircraft acquisition and feet maintenance. Joint purchasing of airpanes and spare parts gives such coectives of buyers greater bargaining power vis-à-vis suppiers. Another strategic innovation, internationa franchising, often aows the franchisee to use the franchiser s airine code and to operate airpanes carrying externa markings of the franchiser. Furthermore, the use of oca franchisees may aso aow an airine to enter the restricted domestic markets of foreign countries. Scheraga (2004) examined the reationship between the strategic focus of airine customer service activities and operationa efficiency (doars per revenue-passenger kiometer). Customer service activities incuded two categories: 1) passenger services, and 2) ticketing, saes and promotion. The resuts of this study suggested that, when these customer service activities distract airines from focusing on those core competencies that aow them to design the operation of their networks of air services from a vaue-based perspective, there are negative impacts on operating efficiency. There is evidence that there is a positive correation between average passenger oad factors and operationa performance (Caves et a. 1981, 1983). Oum and Yu (1999) suggest that average passenger oad factor refects airine s contro of the choice of aircraft and 74 fight frequencies. A higher passenger oad factor indicates better utiization of aircraft and thus it shoud positivey impact operationa efficiency. Finay, severa financia measures of operationa performance are examined. These incude the current ratio, the operating ratio, the net profit margin, the return on investment, and the yied. The current ratio is the ratio of current assets to current iabiities. A ratio of 1.00 is normay considered to refect a sound eve for the airine industry. A ratio ess than 1.00 suggests than an airine may not be generating adequate cash to meet short-term obigations as they become due. A current ratio we above 1.00 suggests an airine may be generating more cash than can profitaby be re-invested for onger-term objectives. At the same time however, such behavior may be reasonabe if the airine is stockpiing cash for potentia acquisitions of other companies or is expecting a future period of aircraft deiveries which wi be bunched together (Morre 2002). The operating ratio is defined as operating revenue expressed as a percentage of operating expenditure. It can be thought of as being simiar to the margin on saes. This ratio gives an indication of management s efficiency in controing costs and increasing revenues. Because this ratio can be distorted by changes in depreciation poicy or a change from ownership of aircraft to operating eases, an aternative form is often used. The aternative definition of the operating ratio is operating profit, after interest charges, expressed as a percentage of operating revenues. Return on investment is the pre-tax profit before interest paid as a percentage of average tota ong-term capita empoyed. This ratio gives an indication of how successfu an airine s management is in its investment of the ong-term capita under its contro. The yied is the revenue generated per revenueton kiometer fown. This ratio captures the impact of an airine s choices with regard to oad management, average stage ength, and passenger mix. The financia and operating data utiized in the study came from the Internationa Civi Aviation Organization s (ICAO) two databases, Financia Data: Commercia Air Carriers, for

7 2003, and Traffic: Commercia Air Carriers, for 2003 and The 2004 data was necessary so that the traffic data coud be carefuy matched to the financia data to aow for the fact that different airines had different fisca year ending dates. A financia data is converted to U.S. doars at rate of exchange which is the average of the 12 month Internationa Transport Association (IATA) rates for the year reported. In those cases where the rate changed consideraby ony in the ast month of the financia year, the rate prevaiing prior to this change was adopted. If the rate for a currency changed frequenty during the year reported, a 12-month average was used. Because exchange rates capture the effects of changes in reative infation and interest rates as we as trade deficits or surpuses, the conversion of a financia vaues to a common currency heps to capture the impacts of idiosyncratic forces on country-specific economies. The ICAO databases are utiized to insure quaity and uniformity of the information provided by airines. There are standardized reporting forms with a reported items carefuy checked by ICAO statisticians. Note that a three Chinese airines were aready issuing equity that was being traded on internationa markets during the period of this study. Such internationa financia activity demanded accuracy and fu discosure of information on the part of these airines. As noted above, an added incentive for accurate financia reporting accuracy was China s desire to enter the Word Trade Organization. METHODOLOGY The methodoogy empoyed is this study is data enveopment anaysis (DEA) and is used to compare the reative efficiencies of the 34 goba fag-carrier airines the sampe. Given a set of inputs and a corresponding set of outputs, a production pan/function is efficient if there is no way to produce more outputs with the same inputs or to produce the same outputs with fewer inputs. In the case of DEA, the production function (and hence the production frontier) is generated from the actua data. DEA is a nonparametric technique that makes no Airine Industry in China assumptions about the form of the production function. Instead, it estimates, in the case of this study, an empirica best practice production frontier from the observed inputs and outputs of the individua airines. The DEA best practice frontier is piecewise inear and approximates the true production function. An is efficient when comparisons with other airines indicate no inefficiency in the utiization of inputs and outputs as measured by its position reative to the efficient production frontier. The anaysis in this study empoys two variations of the two-stage optimization procedure for data enveopment anaysis as specified by Ai and Seiford (1993a). This approach differs from the modes of Charnes, Cooper and Rhodes (CCR) (1978) and Banker, Charnes and Cooper (BCC) (1984). The choice of the two-stage optimization process does not question the vaidity of the CCR and BCC modes. However, the two-stage optimization procedure is a robust computationa methodoogy that reieves the researcher of the need to make discretionary assumptions about some of the underying parameters in these atter two modes. As Ai and Seiford (1993b) note, improper choices in the vaues of these parameters can ead to serious computationa errors. The first variation of this procedure is the base DEA mode. Using the notation of Ai and Seiford (1993a), consider the case of n airines, each utiizing, in varying amounts, m distinct inputs to produce s different outputs. The objective of DEA can be specified so as to minimize tota waste in inputs and outputs. Thus, the objective is: The variabe s r is the amount of sack in, or foregone amount of output r, whie the variabe e i is the excess amount of input i utiized. The vaues and υ r i are shadow prices, or the margina vaue, of a unit of output or input, and λ j is an n x 1 vector of constants utiized in the constraints outined beow and is necessary in the specification of the convexity conditions. The sack variabes aow for the identification of specific inefficiencies in the utiization of 75

8 Airine Industry in China inputs and deficiencies in the generation of outputs. One woud ike the projections and efficiency scores derived to be independent of the units of measurement of the data. To achieve this units-invariant property, the anaysis specified airine specific bounds on the vaues of and υ r i. Given x i the amount of input i used by airine, and y r, the amount of output r for airine, these bounds are defined by: ( 2) 1/, 1,..., 1/, 1,..., µ υ r yr r s and i xi i m The associated resource constraints and convexity conditions, for the base mode aowing for potentia variabe returns-toscae, are discussed in detai in Ai and Seiford (1993a). These are: n ( 3) rjλ j r r 1,..., 1 j n y s y r s 1 1,..., x λ e x i m j ij j i i n ( 4) λ j 1 j 1 ( 5) λ j 0 j ( 6) r 1,..., ( 7) ei 0 i 1,... s r s, m The soution to the above probem identifies, for each airine,, a projected point on the efficient frontier, ( ) where the items x and y are the vectorsof inputs and outputs. The essence of the efficiency evauation of a particuar airine (with an actua achieved combination of x and y ) is the identification of excesses in input utiization (x - ) and deficiencies in output ( - y ). A particuar airine is deemed efficient if (x,y ) =, the airine thus ying on the efficient frontier Thus, one possibe measure of inefficiency, Δ, can be defined by: ( 8) µ ( y y ) υ ( x x) For efficient airines Δ = 0. n The base mode embodies an approach consistent with a systems-oriented phiosophy where input and output variabes are simutaneousy determined. Furthermore, it provides a means of measuring tota inefficiency in both the utiization of inputs and production of outputs. However, to investigate poicy impications for manageria strategies a second mode, the input oriented mode, was aso utiized. The projected point defined above, suggests one possibe path of movement between the observed point and the projected point. Such a movement to the enveopment surface woud be accompished aong the vector (s,-e ). The vector (s,-e ) defines a combination of output augmentation and input reduction. This combination potentiay may consist of two components. The first is a proportiona change in output augmentation and input reduction. The second is a set of additiona (nonproportiona) residua output augmentation and input reduction after the initia proportiona changes have taken pace. This can be written as foows: Thus for each airine, the output vector can ( 9) s y ρ δ s and e γ x δe be increased proportionatey (in each vector component) by a factor of ρ with individua non-proportiona residua component increases in each of the separate output variabes given by δ s. Simiary, the input vector for an airine can be decreased proportionatey (in each vector component) by a factor of γ with individua non-proportiona residua component decreases in each of the separate input variabes given by δ e. Note that at east one eement of the δ and s δ vectors is zero. e In the case of the input oriented mode, as opposed to the base mode, one set of variabes, inputs, takes priority over the other. For the input orientation mode, proportiona movement to the enveopment surface from the observed point is first achieved in input-space. That is, the mode seeks a projected point so that γ is maximized. Because this mode requires, in the first stage, the maximization of γ, it effectivey identifies the intermediate point [(1-γ)x, y ]. Letting θ = 1-γ, for the input oriented mode the 76

9 input constraint is now repaced by: n ( 10) ijλ j θ i δi 0 1,..., x x i m 1 j Notice, that effectivey, the input oriented mode requires the soution of the inear program with regard to the intermediate point [(1-γ)x,y ]. The projected point obtained with this orientation can differ from that obtained from the nonoriented base mode. However, the projected point wi sti ie on the enveopment surface and Δ = 0. The input oriented mode seeks a projected point such that the proportiona reduction in inputs is maximized. The impicit underying premise in such an orientation is that the primary objective of the airine under evauation is to gain efficiency by reducing excess input utiization whie continuing to operate with its current technoogy mix (refected in actua input ratios). As noted above, satisfaction of the primary or first-stage objective may not be sufficient for the attainment of the enveopment surface and hence efficiency. In genera, in the second-stage, a non-oriented projection has to be appied to the intermediate point. The input-oriented mode generates another vauabe piece of information. In addition to the maximization of γ, a measure of efficiency for the input oriented mode is generated which is reated to the weighted-ratios of the CCR mode (1978) and defined as: Input efficiency, ι, is the mutipe of the input ( 11) i 1 I where I υ x vector that woud pace the reativey inefficient airine on the efficient production frontier. Effectivey, ι measures the tota inefficiency in terms of proportiona input reduction. Aso note, that, for an efficient airine, ι = 1. Thus, if ι takes on the vaue of 0.75 for a particuar airine, this is equivaent to saying that this reativey inefficient airine coud reduce a inputs by the proportion , or 25% and move to the efficient production frontier. The efficiency measure ι does convey information with regard to manageria poicy. Consider the foowing case. Suppose airine A i Airine Industry in China has a peer group of airines with comparativey efficient production techniques aowing them to achieve the eves of output of airine A more efficienty. If ι is very sma, then the production technoogy of airine A is very bady chosen. The airine woud be prudent to shift the input/production technoogy with a focus on increasing eves of output. If, on the other hand, ι is cose to 1, then the airine coud remain with its current production technoogy and achieve the same eves of output with a sma scaing down of inputs. Thus, utiization of the input-oriented mode in conjunction with the base mode aows the researcher to not ony deveop assessment measures of inefficiency but aso to evauate the efficacy of manageria strategies. The software utiized for the data enveopment anaysis was an updated version of that utiized by Schefczyk (1993), Integrated Data Enveopment Anaysis System, Version obtained from 1 Consuting. In investigating the impact of the above operationa and environmenta variabes on reative operationa efficiency, note that ι is a censured variabe, i.e. 0 ι 1. Utiization of ordinary east squares yieds biased and inconsistent parameter estimates (Pindyck and Rubinfed 1991, p. 276). The more appropriate tobit technique is that deveoped by Tobin (1958) for a eft-censured variabe. A transformation suggested by Fethi et a. (2002) is utiized to transform ι into a eftcensured variabe. This new variabe is defined as (1/ ι ) 1 which is greater than or equa to zero in a continuous fashion. Thus, for the transformed vaue of ι, an efficient airine wi have a vaue of zero, whie inefficient airine wi have a vaue greater than zero. The tobit anaysis was performed using the LIFEREG procedure in the SAS statistica package. RESULTS As shown in Tabe 2, the input-oriented mode generated vaues for ι for Air China, China Eastern Airines, and China Southern Airines that are respectivey 0.86, 0.71, and Whie Air China is one-haf of a standard deviation beow the mean vaue of ι, China Eastern Airines is two standard deviations beow the 77

10 Airine Industry in China Tabe 2: Sampe Input-Oriented Operating Efficiencies NAME ι Aeroines Argentinas 1.00 Aeromexico 0.94 Air Canada 0.84 Air China 0.86 Air France 1.00 Air India 0.95 A Nippon Airways 0.73 American Airines 1.00 Asiana Airines 0.91 Austrian Airine Group 0.92 British Airways 0.89 Cathay Pacific 1.00 China Eastern Airines 0.71 China Southern Airines 0.58 Continenta Airines 1.00 CSA Czech Airines 1.00 Deta Airines 0.89 EL AL 1.00 Iberia 0.87 KLM 1.00 Korean Air 0.76 LOT Poish Airines 0.86 Lufthansa 1.00 Maaysia Airines 1.00 Mexicana 0.81 Northwest Airines 0.90 Pakistan Internationa Airines 0.90 Scandinavian Airines 0.85 Singapore Airines 1.00 SriLankian Airines 1.00 TAP Air Portuga 0.91 Tarom 1.00 Thai Airways Internationa 1.00 United Airines 1.00 Mean Std. Dev. Min. Max. ι

11 mean and China Southern Airines is more than three standard deviations beow the mean. In fact, China Eastern Airines and China Southern Airines demonstrated the two owest vaues for ι in the sampe of 34 airines. It is tempting to concude that an exogenous phenomenon such as the SARS outbreak in China is the driver of these resuts. Reca, however, that as noted above, a very sma vaue for ι is suggestive of the need for an airine to reconsider the input configuration that it utiizing to produce a given set of outputs. This can be further seen by examining the resuts of the base mode for the three Chinese airines in Tabe 3. In the case of a three Chinese airines, there is no inefficiency with regard to the input of avaiabe capacity as measured by avaiabe ton-kiometers. Furthermore, in the case of Air China, there is no inefficiency with regard to the output of revenue passenger-kiometers. For a three airines, the argest output inefficiencies are with regard to non-passenger ton-kiometers. As shown in Tabe 4a, for the entire sampe, the average vaue for the percentage of revenues from passenger services was 80.5%. The respective vaues (from Tabe 4b) for Air China, China Eastern Airines, and China Southern Airines were 73, 76, and 89%. Thus, at east for former two airines, a significant portion of their revenues are generated by freight and mai services. Furthermore, Tabe 3 indicates that a three airines demonstrated very arge inefficiencies with regard to operating cost (refecting operating cost excuding capita and aircraft cost refected in avaiabe tonkiometers) and non-fight assets (e.g. faciities, reservation systems, current assets). Airine Industry in China Note that operating costs incude fue cost. On average, fue can constitute 14 16% of operating costs. For some of the major airines in the industry, this may be as much as 20%. Furthermore, shorter hau airines typicay get ower fue efficiency because take-offs and andings consume high amounts of jet fue. However, it must aso be noted that rising fue costs were not unique to the three Chinese airines. Tabes 5a and 5b present the tobit regressions resuts describing the impact of the operationa and environmenta variabes discussed above. Reca, that for the transformed vaue of ι, an efficient airine wi have a vaue of zero, whie inefficient airine wi have a vaue greater than zero. Thus variabes positivey correated with ι wi be negativey correated with the transformed vaue of ι. Except, for expenditures on passenger services revenue passenger-kiometer, a of the operationa and environmenta variabes are statisticay significant at the 10% eve or better. Tabe 5a demonstrates that except for the percentage of revenues from schedued operations, the signs on the coefficients of the variabes are ogicay intuitive. The impied negative impact of an increase in the percentage of revenues from schedued operations is, at first, somewhat puzzing. However, a statisticay significant correation was found between this variabe and the percentage of revenues from passenger services. Therefore, an interaction term between these two variabes was introduced into the regression. Tabe 5b demonstrates a more reasonabe scenario as the interaction variabe is positive. An increase in the percentage of Air China (CCA), China Eastern Airines (CES), China Southern Airines (CSN) CCA (%) CES (%) CSN (%) Avaiabe Ton-Km Operating Cost Non-Fight Assets Revenue Passenger-Km Non-Passenger Ton-Km Negative/Positive Sign = Necessary Reduction/Increase Required to Achieve Efficiency Tabe 3: Input and Output Percentage Inefficiencies 79

12 Airine Industry in China Tabe 4a: Sampe Descriptive Statistics: Tobit Regression Variabes DependentVariabe: Transformed Iota: [(1/ι) 1] Variabe Mean Std. Dev. Min. Max. Percentage of Revenues from Internationa Operations 73.03% 28.05% 19.00% 100% Percentage of Revenues from Schedued Operations 88.62% 7.40% 68.00% 100% Percentage of Revenues from Passenger Services 80.50% 8.82% 56.00% 94.00% Average Stage-Length 1, Km Km Km 5, Km Passenger Load Factor 71.32% 4.96% 60.00% 80.00% Expenditures on Passenger Services per Revenue Passenger-Km $ $ $ $ Expenditures on Ticketing Saes and Promotion per Revenue Passenger-Km $ $ $ $ Tabe 4b: Descriptive Statistics: Tobit Regression Variabes DependentVariabe: Transformed Iota: [(1/ι ) 1] Variabe CCA CES CSN 38% (-1.25) Percentage of Revenues from Internationa Operations Percentage of Revenues from Schedued 89% Operations (0.05) Percentage of Revenues from Passenger 73% Services (-0.85) Average Stage-Length 1, Km (-0.19) Passenger Load Factor 66% (-1.07) Expenditures on Services per Revenue Passenger-Km Expenditures on Ticketing Saes and Promotion per Revenue Passenger-Km $ (-2.84) $ (-0.80) 29 (-1.57) 94% (0.73) 76% (-0.51) 1, Km (-0.51) 61% (-2.08) $ (-2.66) $ (-0.22) 19% (-1.93) 99% (1.40) 89% (0.96%) 1, Km (-0.58) 64% (-1.48) $ (-2.62) $ (0.22) (Figures in parentheses represent number of standard deviations above or beow the mean for the entire sampe.) 80

13 Airine Industry in China Tabe 5a: Tobit Regression Resuts DependentVariabe: Transformed ι Variabe Estimate Chi-Square Pr>ChiSq Intercept Percentage of Revenues from Internationa Operations Percentage of Revenues from Schedued Operations Percentage of Revenues from Passenger Services Average Stage-Length Passenger Load Factor Expenditures on Services per Revenue Passenger-Km Expenditures on Ticketing Saes and Promotion per Revenue Passenger-Km Tabe 5b: Regression Resuts with Interaction Term DependentVariabe: Transformed ι Variabe Estimate Chi-Square Pr>ChiSq Intercept Percentage of Revenues from Internationa Operations Percentage of Revenues from Schedued Operations Percentage of Revenues from Passenger Services Percentage of Revenues from Schedued Operations X Percentage of Revenues from Passenger Services Average Stage-Length Passenger Load Factor Expenditures on Services per Revenue Passenger-Km Expenditures on Ticketing Saes and Promotion per Revenue Passenger-Km

14 Airine Industry in China revenues from schedued operations, per se, has a positive impact on reative operationa efficiency. However, airines in the sampe that had higher vaues for this variabe aso tended to have higher percentages of revenues from passenger services, that is, higher vaues of schedued passenger services. This, then, had a negative impact operationa efficiency. This is aso consistent with the operationa advantages of carrying cargo suggested by O Connor (2001) and noted above. These advantages are not trivia in the sampe of this study as Tabe 4a indicates about 20% of revenues, on average, were generated from non-passenger services (freight and mai). The resuts reported in Tabe 5b and the information in Tabes 4a and 4b aow for a discussion of the impact of the operationa and environmenta variabes on Air China, China Eastern Airines, and China Southern Airines. A three airines are significanty ower than the sampe average with regard to the percentage of revenues from internationa operations. Thus they are not capturing the positive impacts on operationa efficiency of goba strategic aiances, biatera agreements, and internationa franchising. The three Chinese airines are significanty ower than the sampe average with regard to passenger oad factors. Some of this certainy can be attributed to the SARS epidemic and, as suggested by the regression resuts, wi negativey impact reative operationa efficiency. However, as noted above, to abe this factor as the overwheming determinant of the reative operationa inefficiency of the three Chinese airines woud be simpistic and 82 miseading. A three Chinese airines are somewhat beow the sampe average with regard to average stage ength. Thus, as compared to other airines in the sampe, they are not as effectivey taking advantage of economies of distance, which drive an inverse reationship between average fight ength and unit cost a reationship that enhances operationa efficiency. Air China and China Eastern Airines are beow the sampe mean for the percentage of revenues from passenger services but above the sampe mean for percentage of revenues from schedued services. Given the statisticay significant impact of interaction term discussed above, shoud they decide to increase the percentage of revenues from schedued passenger services, they wi have to be carefu in ensuring that the necessary resources and systems are in pace to maintain or enhance operationa efficiency. Additionay, Air China and China Eastern Airines are beow the sampe mean for expenditures on ticketing saes and promotion per revenue passenger-kiometer. However, as the regression resuts demonstrate, these expenditures tend to negativey impact operationa efficiency. Thus, the behavior of these two airines may be economicay reasonabe as compared to China Southern Airines that is above the sampe mean. Tabes 6a and 6b aow for a comparison of the three Chinese airines reative to the entire sampe with regard to severa performance measures. For the current ratio, a three airines are significanty beow the sampe mean of 0.82, et aone the desired vaue of Air China, China Eastern Airines, and China Tabe 6a: Sampe Descriptive Statistics: Performance Measures Variabe Mean Std. Dev. Min. Max. Current Ratio Operating Ratio Aternative Operating Ratio Net Profit Margin Return on Investment Yied

15 Tabe 6b: Individua Airine Performance Measures Airine Industry in China Current Oper. At. Oper. Net Prof. NAME Ratio Ratio Ratio Margin ROI Yied Aeroines Argentinas Aeromexico Air Canada Air China Air France Air India A Nippon Airways American Airines Asiana Airines Austrian Airine Group British Airways Cathay Pacific China Eastern Airines China Southern Airines Continenta Airines CSA Czech Airines Deta Airines EL AL Iberia KLM Korean Air LOT Poish Airines Lufthansa Maaysia Airines Mexicana Northwest Airines Pakistan Internationa Airines Scandinavian Airines Singapore Airines SriLankian Airines TAP Air Portuga Tarom Thai Airways Internationa United Airines Southern Airines had vaues of 0.45, 0.52, and 0.52 respectivey. Thus, they are certainy not generating adequate cash to meet short-term obigations. The three airines are above the sampe mean for the operating ratio measure. In this regard, they are better than a significant number of airines in the sampe with regard to their management s efficiency in controing costs and increasing revenues. For the aternative specification of this measure, Air China is we above the sampe average, whie China Eastern Airines and China Southern Airines are sighty beow it. There is no consistent pattern for the Chinese airines on the net profit margin and return on investment measures. Air China is above the sampe mean with regard to 83

16 Airine Industry in China net profit margin; China Southern Airines is at the mean; and China Eastern Airines is beow it. However, China Eastern Airines and China Southern Airines are above the sampe mean with regard to return on investment whie Air China is beow it. Most interestingy, a three airines are beow the sampe mean vaue for the yied measure. This woud seem to be consistent with the ow vaues of ι as we as the vaues for some of the operationa and environmenta variabes for these airines. The yied measure, as noted above, refects choices with regard to oad management, average stage ength, and passenger mix. CONCLUSION It is interesting to examine the chairman s statement in each of the 2004 (a year after this study) annua reports for Air China (2005), China Eastern Airines (2005), and China Southern Airines (2005). There are some consistent patterns in perceived operationa needs needs highighted by the anaysis of this study. The Air China annua report highights the need for functiona integration to transform and streamine the organizationa structure to enhance operations efficiency. The China Southern Airines annua report describes the adoption of measures to optimize service offerings, fight schedues and route networks by improving integrated management and resources aocation. In a simiar manner, the 84 China Eastern Airines annua report highights the need to beef up the management of fight equipment in a comprehensive manner. A three airines focus on the importance of cargo operations. Air China sees the need to depoy more air cargo capacity whie optimizing the cargo transportation network. China Southern Airines seeks to strengthen the cargo operation to further penetrate internationa and regiona markets. China Eastern Airines sees itsef continuing the strategy for deveoping the cargo transport business and promoting the deveopment of aviation ogistics. Finay, a three airines perceive a need for better budget management. Wang Yongtao (2002), an industry researcher with Air China, has emphasized that a key issue in the Air China consoidation is management. He noted that if the consoidation (Air China Group) just combines the assets of the three airines and acks improvement in management, the reform wi achieve no effective resut. Further, he suggests that ony when the new group pays enough attention to efficiency and market demands, can the air group expect a better future. This preiminary study has sought to identify specific areas of operations management practices, reative to a arge comparative set of goba fag carriers, that need to be addressed in order to achieve operationa efficiency. References Air China Ltd., 2004 Annua Report, Beijing, Peope s Repubic of China, Apri, Ai, A. and L. Seiford. The Mathematica Programming Approach to Efficiency Anaysis. H. Fried, C. Love, and S. Schmidt eds. The Measurement of Productive Efficiency. New York: Oxford University Press (1993a): Ai, A. and L. Seiford. Computationa Accuracy and Infinitesimas in Data Enveopment Anaysis. INFOR 31 (4), (1993b): Banker, R.D., A. Charnes, and W.W. Cooper. Some Modes for Estimating Technica and Scae Inefficiencies in Data Enveopment Anaysis. Management Science 30 (9), (1984): Banker, R. D. and H. H. Johnston. An Empirica Study of Cost Drivers in the US Airine Industry. The Accounting Review Juy, (1993): Caves, D. W., L. R. Christensen, and M. W. Trethaway. US Trunk Air Carriers : A Mutiatera Comparison of Tota Factor Productivity. T. G. Cowing and R. E. Stevenson eds. Productivity Measurement in Reguated Industries. New York: Academic Press (1981): Caves, D. W., L. R. Christensen, and M. W. Trethaway. Productivity Performance of US Trunk

17 Airine Industry in China and Loca Service Airines in the Era of Dereguation. Economic Inquiry 21, (1983): Caves, D. W., L. R. Christensen, and M. W. Trethaway. Economies of Density versus Economies of Scae: Why Trunk and Loca Service Airines Costs Differ. Rand Journa of Economics Winter, (1984): Charnes, A., W. Cooper, and E. Rhodes. Measuring the Efficiency of Decision Making Units. European Journa of Operationa Research 2 (6), (1978): China Eastern Airines Corporation Ltd., 2004 Annua Report, Shanghai, Peope s Repubic of China, Apri, China Southern Airines Company Ltd., 2004 Annua Report, Guangzhou, Peope s Repubic of China, Apri, Dana, L. and T. Dana. Internationa Airine Connections. ANZMAC98 Conference Proceedings. Dunedin, New Zeaand: University of Otago (1998): Efendiogu, A. and W. Murray. Changes and Chaenges in the Airine Industry in China. Proceedings of the Seventh Internationa Conference on Goba Business and Economic Deveopment. Bangkok, Thaiand (2003): Fethi, M., P. Jackson, and T. Weyman-Jones. Measuring the Efficiency of European Airines: An Appication of Tobit Anaysis. Working Paper, University of Leicester Management Center. Leicester: United Kingdom (2002). Internationa Civi Aviation Organization. Financia Data: Commercia Air Carriers. Montrea, Quebec, Canada, Internationa Civi Aviation Organization. Traffic: Commercia Air Carriers. Montrea, Quebec, Canada, 2003 and Morre, P. Airine Finance, 2 edition. Hants, Engand: Ashgate Pubishing Ltd, O Connor, W. E. nd An Introduction to Airine Economics. Westport, Connecticut: Praeger, Consuting. Integrated Data Enveopment Anaysis System, Version Amherst, Massachusetts, Oum, T. and C. Yu. Winning Airines: Productivity and Cost Competitiveness of the Word s Major Airines. Norwe, Massachusetts: Kuwer Academic Pubishers, Pindyck, R. and D. Rubinfed. Econometric Modes & Economic Forecasts. New York: McGraw Hi Inc, Schefczyk, M. Operationa Performance of Airines: An Extension of Traditiona Performance Paradigms. Strategic Management Journa 14, (1993): Scheraga, C. The Reationship between Operationa Efficiency and Customer Service: A Goba Study of Thirty-Eight Large Internationa Airines. Transportation Journa 43 (3), (2004): Siau, C. and D. Van Lindt. Cost Driver Anaysis in the Airine Industry. Journa of Cost Management 11, (1997): Tobin, J. Estimation of Reationships for Limited Dependent Variabes. Econometrica 26, (1958): Trethaway, M. An Internationa Comparison of Airines. Proceedings of the Canadian Transportation Research Forum, (1984): Yongtao, W. quoted in China s Domestic Air Industry Set to Undergo Consoidation. Civi Aviation Resource Net of China (CARNOC) 3 (March), (2002).! " # $ % is professor of business strategy and technoogy management. His fieds of research 85

18 Airine Industry in China and teaching incude transportation and internationa ogistics, goba strategic management, cross-cutura management, and the management of technoogy and innovation. Scheraga received his Ph.D. in economics from the University of Connecticut, an M.A. in economics, and a Sc.B. in mathematics and engineering from Brown University. 86