vic

Department of Business Administration, National Taipei College of Business, 321, 1, Jinan Rd

nciia, aperratithe

e carristomer

A major perceived area of cost savings is ticosts, but there are also likely potential sa

d elimABC) sl costsllocati. In th

practices in industrial organizations (Anderson, 1995; Compton,1996).1 In a business organization, the ABC methodology assignsan organizations resource costs through activities to the products

the performance of international airlines. While there are alterna-tive techniques, generally involving some form of stochastic

E-mail address: wencheng@webmail.ntcb.edu.tw.1 As explained by Peter F. Drucker (1999), traditional cost accounting focuses on

what it costs to do something, for example, to cut a screw thread; activity-basedcosting also records the cost of not doing, such as the cost of waiting fora needed part. In this way, activity-based costing records the costs that traditional

2 ABC is not without its limitations, and in particular, manually driven ABC canbean inefcient use of resources; it can be expensive and difcult to implement forsmall gains, but running against this it covers a broader-broad band of activitiesthan many alternative.3 Cooper et al. (1999) is a comprehensive text that explains DEA models and

applications and provides further reading as well as DEA-Solver software.

Contents lists available at SciVerse ScienceDirect

Journal of Air Transport Management

journal homepage: www.elsevier .com/locate/ ja ir t raman

Journal of Air Transport Management 19 (2012) 13e15cost accounting does not do.relationship between nancial performance and customer serviceusing data envelopment analysis (DEA) as part of the ABC process.

2. Methodology

The ABC system, which has its basis in activity costing andinputeoutput accounting, has developed over the past 20 years asa tool for improving the behavioral, business and accounting

associated with activities, ABC can also evaluate whether thoseactivities add value, thus providing away to assess efciency and/orenhance customer services. It also has the practical appeal to manyrms that it is focused on activities, not responsibilities, and so isseen as less threatening to the mangers of the various intra-business functions under review.2

To examine this efciency aspect within an ABC context wedeploy a long established, and well-tried programming technique;data envelopment analysis3 as part of an ABC process in assessingbut also seeks to improve inefciencyimproved operational processes anciencies. The activity-based costing (way to help identify ways to controseries. It is not only concerned with a0969-6997/$ e see front matter 2011 Elsevier Ltd.doi:10.1016/j.jairtraman.2011.12.002services that they offer.me in transit and ightvings to be made fromination of other inef-ystems offer a potentialand enhance customerng costs more preciselyis paper, we explore the

In more detail, the approach initially involves dividing theproduction procedure into a series of activities and allocatingoverhead costs to each. This model assigns more indirect costs intodirect costs compared to conventional costing models. Then, basedon the levels of these activities consumed by the nal products orservices being produced, it allocates overhead cost to each of these.Production costs are thus allocated through a cost driver methodin two stages to minimize distortions. In focusing on the coststhe airline industry, has caused largthey operate and the nature of the cuers to reassess the way process improvement initiatives.

bilateral air service agreement, and the repeated nancial crises in pricing, outsourcing, and the identication and measurement ofKeywords:Financial performance of airlinesAirline customer serviceActivity-based costing for airlines

a b s t r a c t

This study looks at the naEurope, Latin America, Asfocused expenditures on oanalysis, indicate that opemay be improved through

1. Introduction

The trend towards increasingly commercially drivenmarkets forinternational air services accompanying the liberalization of manyFinancial performance and customer sercosting of 38 international airlines

Wen-Cheng LinAll rights reserved.e: An examination using activity-based

., Taipei City, Taiwan, ROC

al performance of a set of large international airlines from North America,nd the Middle East. Efciency measures are related to their strategicallyations and on customer services. The results, based on data envelopmenton management, including that of customer service attribute evaluation,adoption of activity-based costing analysis.

2011 Elsevier Ltd. All rights reserved.

and the services that it provides to its customers. It is generally usedas a tool for understanding product and customer cost and prot-ability based on the production or performing processes. As such,ABC has largely been used to support strategic decisions such as

by Caves et al. (1983) and others have shown evidence of a positive

2.2. Stage 2: data envelopment analysis

Our use of ABC involves generating data on cost drivers forseparate services and transactions that are then fed into a DEA tomeasure the relative efciency of each. For the input-orientedmodel, efciency measures the proportions of inputs used bya relatively efcient airlines to produce its current level of output,with the most efcient carriers having values of one. DEA analysisdoes not provide guidance as to the potentially most efcient use ofresources but rather it develops as benchmarks a list of airlines thatare at the frontier of efciency of those in the market. It may bepossible to be more efcient than these, but no current airline, or atleast none in the data set, is at that level.

Table 1Airline efciency values.

Airline Country Efciency value

British Airways United Kingdom 0.99Virgin Atlantic United Kingdom 1.00American Airlines United States 1.00Continental Airlines United States 1.00Delta Airlines United States 1.00Northwest Airlines United States 1.00Trans World Airline United States 0.94United Airlines United States 1.00US Airways United States 0.82Aerolineas Argentinas Argentina 1.00Austrian Airlines Austria 0.86Biman Bangladesh Bangladesh 0.91Lloyd Aereo Boliviano Bolivia 1.00

LOT Poland 0.86Asiana Republic of Korea 1.00Korean Air Republic of Korea 1.00Scandinavian Scandinavia 0.77Singapore Airlines Singapore 1.00Iberia Spain 0.91Srilankian Sri Lanka 0.99Thai Airways International Thailand 1.00

Table 2Mean values of Tobit regression variables.

Variables Mean Minimum Maximum

International percentageof operations

73.35 13.238 100

Load factor: ton-kilometers (%) 60.39 42.34 82.53Percentage of operating

revenues from passenger services (%)81.12 58 99

Scheduled service revenuesas a percentage of revenues (%)

91.74 79 99

Operations focus ($/ATKm) 0.25 0.178 0.728Ticketing, sales, promotion

focus ($/revenue passengers)0.04 0.012 0.072

Indirect cost ($thousands) 268 12 1589Fixed assets ($thousands) 3571 123 56,854

ort Management 19 (2012) 13e15correlation between average load factors and nancial performancewhile Oum and Yus work (1999) suggests that average load factoris a reection of an airlines choice of aircraft and ight frequencies.Fethi et al. (2002) nd that an increase in the international focus ofan airline leads to spatial disparities in its nancial environment.The share of scheduled service revenues is anticipated to havea positive impact on operational efciency; scheduled ightsrequire different products and marketing than charter ights thatallows a rationalization of routines and greater overall efciency.

In terms of passenger/cargo mix, Oum and Yu (1999) argue thatthis can be particularly important in Asian and European marketswhere the latter accounts for a large portion of output of manyAsian and European carriers based in export-oriented countries.The economic of carrying passenger and cargo is different; cargo forexample is usually carried one-way, while passengers generallymake round trips. Average ight length captures economies ofdistance that Trethaway (1984) and others have demonstratedaffects costs.

Activity-based cost accounting breaks down cost estimation intofour stages:

identify activities, assign resource costs to activities, identify outputs, link activity costs to outputs.

ABC differs from traditional cost accounting that assigns costsdirectly to outputs (Cooper and Kaplan, 1992). The information oncosts, activities, and outputs that ABC generates can be used in DEAthat seeks individual producing units that are getting the mostoutput of their inputs for each activity or prot center. TogetherABC and DEA can provide a two-dimensional portrayal of a businessacross individual operating units and individual inputs and outputs.

2.1. Stage 1: activity-based costing

We distinguished between two basic activities: providingservices for passengers and carrying out transactions. Dis-tinguishing between these is the basis of any carriers route plan-ning; essentially projecting what market conditions on variousroutes will look like in a decade or so and to then preparing forthese changes well in advance. One aspect of the ABC approach is toemphasizes drivers; factors that drive costs and that drive income.There is a major difference between cost drivers for providingservices and cost drivers for carrying out transactions with regardto a route. Providing services for customers involves providingthem with access to ight transactions anywhere in the countryusing a variety of technologies; whereas route services arefrontier analysis, the DEA methodology avoids the need for aninitial specication of a functional form for the efciency frontier.The basic idea of DEA, as with stochastic approaches, is to identifythe most efcient decision-making unit (DMU). In this context,efciency is the ratio of the weighted sum of a rms outputs to theweighted sum of its inputs. Once the efciency scores are derived,they are regressed against a set of nancial and operating variablesthat include the percentage of passenger operations, passengerrevenue at the average load factor, international passengerrevenue-kilometers as a percentage of passenger revenue-kilometers, scheduled service revenues as a percentage of reve-nues, sales, indirect cost, xed assets, and receivable turnover indays.

In terms of anticipated relationships and expected signs, studies

W.-C. Lin / Journal of Air Transp14geographically specic.Air Canada Canada 0.85Canadian Airlines Canada 0.84Lan Chile Chile 1.00Cathay Pacic China, Hong Kong 1.00Avianca Colombia Colombia 0.75Czech Airlines Czech Republic 1.00Air France France 1.00Lufthansa Germany 0.97Air India India 0.91All Nippon Airways Japan 0.73Japan Airlines Japan 1.00Kuwait Airways Kuwait 0.86Malaysia Airlines Malaysia 0.94Aeromexico Mexico 0.93Mexicana Mexico 0.91KLM Royal Dutch Airlines Netherlands 1.00Pakistan International Airlines Pakistan 0.84Receivable turnover (days) 28 10 58

interactive terms are included. These involve multiplying each ofthe nancial variables by the ABC measures. The independentvariables are also standardized to limit multicollinearity problems.

Five of the variables examined are seen to be signicant at the10% level or better. Focusing initially on the direct effects, in linewith previous studies load factor has the expected positive impacton nancial efciency (costs decline as load-factor rises) and issignicant. Similarly, there is a signicant positive effect on costs

Table 3Tobit model results.

Intercept 12,485Indirect cost ($ thousands) 1.391International percentage of operations 0.385**Load factor (%) 1.259***Percentage of operating revenues from passenger services 0.956**Scheduled service revenues as a percentage of revenues 1.676*Operations focus ($/ATKm) 0.463

W.-C. Lin / Journal of Air Transport Management 19 (2012) 13e15 15We use this approach to develop benchmarks against whichairlines can compare their performance; essentially their relativeefciency. The aim is to provide guidance for less efcient carriersas to way of increasing their productivity and more closely timetheir activities to consumer demands.

3. Results

To make inter-rm comparisons meaningful, it is necessary tocompare rms that are similar regarding such things as the busi-ness environment they operate in, their mix of activities, and theirsize. It makes no sense, for example to compare a large interna-tional carrier such as Lufthansa with a small regional airline. To thisend Kantor and Maital (1999) conducted a cluster analysis ofairlines that involved dividing them into groups that have a highdegree of internally homogeneous. We use the grouping developedin our applications of DEA, of nancial statement analysis, and ofABC to international airlines.

We use data for thirty-eight international airline for scal year2008 taken from the International Air Transport Association,International Civil Aviation Association and nancial statement ofthe carriers. The results of the DEA analysis are seen in Table 1,where we see that efciency values range from a low of 0.73 toa high of 1.00. Thirty4 of the 38 airlines in the sample are seen to beoperating with the efcient frontier. In terms of interpretation, All

Ticketing, sales, promotion focus ($/revenue passengers) 6.422Fixed assets ($ thousands) 3.158Receivable turnover (days) 0.549Indirect cost international percentage of operations 0.151Indirect cost load factor 1.448Indirect cost percentage of operating

revenues from passenger services2.763***

Indirect cost scheduled servicerevenues as a percentage of revenues

1.612

Signicance levels using Chi-squaredtests: *p < 0.1; **p < 0.05; ***p < 0.01.Nippon Airways with an efciency value of 0.74, could havereduced all inputs proportionately by 26% and still have producedthe same level of output. Other airline that could have reduced allinputs proportionately by more than 10% include US Airways,Austrian Airlines, Air Canada, Canadian Airlines, Avianca Colombia,Kuwait Airways, Pakistan International Airlines, LOT andScandinavian.

Table 2 presents the mean values of the explanatory variablesused in the subsequent Tobit analysis, the results of which,including statistical signicance, are presented in Table 3.5 Inaddition to just using variables independently, a number of

4 There is a tendency when using DEA to have a larger proportion of observationson the frontier as the sample size increases and as more inputs are considered. Thisis because each individual observation e airlines in this case e tends to have morespecic features making it the most efcient at offering that particular bundle ofoutputs using that collection of inputs. The number on the frontier in thesecalculations is within the range one would expect for the type of analysis beingconducted here.

5 A Tobit or similar non-linear sigmoid model is required because the values fromthe DEA analysis are bounded by one and zero.from offering a high percentage of scheduled services. It is alsoseen, however, that a specialization in serving international, andpassenger markets exerts a signicant negative impact on ef-ciency; costs rise as they increase. This is also true of the interactionterm between the indirect cost and the percentage of operatingrevenues from passenger services.

4. Conclusions

Traditional cost accounting methods, which allocate carriersindirect costs on the basis of one driver, such as ight length oraircraft load factor, while simple to do and understand, canexcessively narrow and ultimately be misleading. An effectiveapplication of ABC procedures can incorporate a variety of poten-tially important factors that inuence airline performance. Here weprovide a case study by using DEA to conduct a nancial statementanalysis of 38 large international airlines. The results show theimportance of operating scheduled services on lowering costs, andof having a high load factor.

Acknowledgments

This work was supported by National Science Council, Taiwan,ROC under grant NSC 99-2410-H-141-013. The authors would liketo thank anonymous reviewers for their helpful comments.

References

Anderson, S., 1995. A framework for assessing cost management system changes:the case of activity-based costing implementation at general motors. Journal ofManagement Accounting Research 7, 1e51.

Caves, D.W., Christensen, L.R., Trethaway, M.W., 1983. Productivity performance ofUS trunk and local service airlines in the era of deregulation. Economic Inquiry21, 312e324.

Compton, T.R., 1996. Implementing activity-based costing. The CPA Journal 66,20e27.

Cooper, R., Kaplan, R.S., 1992. Implementing Activity-Based Cost Management.Prentice Hall, Montvale.

Cooper, W.W., Seiford, L.M., Tone, K., 1999. Data Envelopment Analysis: AComprehensive Text with Models, Applications, References and DEA-SolverSoftware. Kluwer, Boston.

Drucker, P., 1999. Management Challenges of the 21st Century. Harper Business,New York.

Fethi, M., Jackson, P., Weyman-Jones, T., 2002. Measuring the Efciency of EuropeanAirlines: An Application of Tobit Analysis. Working Paper. University ofLeicester, Management Center.

Kantor, J., Maital, S., 1999. Measuring efciency by product group: integrating DEAwith activity-based accounting in a large mideast bank. Interfaces 29, 27e36.

Oum, T., Yu, C., 1999. Winning Airlines: Productivity and Cost Competitiveness ofthe Worlds Major Airlines. Kluwer Academic Publishers, Norwell.

Trethaway, M., 1984. An international comparison of airlines. Proceedings of theCanadian Transportation Research Forum, 34e43.

Financial performance and customer service: An examination using activity-based costing of 38 international airlines1. Introduction2. Methodology2.1. Stage 1: activity-based costing2.2. Stage 2: data envelopment analysis

3. Results4. ConclusionsAcknowledgmentsReferences