Physical Productivity versus Financial Efficiency: a Case Study of Major Airports in the Asia Pacific Region

Transcription

1 World Journal of Management Vol. 6. No. 2. September 2015 Issue. Pp Physical Productivity versus Financial Efficiency: a Case Study of Major Airports in the Asia Pacific Region JEL Codes: D24, L93, M11, R40 1. Introduction Tae Seung Kim This study tries to figure out the difference between the physical productivity and the financial efficiency among major airports in the Asia-Pacific region using the data of 11 airports for the period Based on the super- SBM, 4 different efficiency scores are measured and the origins of the differences of them are identified through the calculation of correlation coefficients. Another findings are that the market in the region is led by the airports in China and neighbors in terms of efficiency, and that the efficiency in the region is getting improved over time. Recently, efficiency becomes one of the most important performance indicators in operation not only of private companies but also of public entities. Due to the severe competitive environment in the world economy especially after globalization from 1980 s, private companies have had to seek their sources of profits from the optimization of operation in the supply side as well as from the expansion or differentiation of markets in the demand side. The continual trends of privatization, hence the introduction of the market competition system, elaborated by central/local governments, have urged the public corporations to strive for managerial reformation to cope with the imminent competition with the private sectors. Under those circumstances, numerous indicators have been developed to capture the performances of firms from various standpoints. Aside from quantitative indices, such as sales, market share, growth rate, and indices for accounting profitability, several qualitative reference marks have been adopted in many industry sectors, among which efficiency indices must have attracted more attentions of executive managers in related companies as well as scholars in the fields. Efficiency can be interpreted as the productivity in terms of physical volumes of inputs and outputs, and as the financial efficiency in terms of monetary inputs and outputs. These two aspects of efficiency can be explained by the theory of duality in economics in which the optimized point of the production function in a firm is the same with that of the cost function in the firm, if the monetary inputs and outputs are measured by the input prices and the total costs, respectively, in the financial efficiency. address: todang@inha.ac.kr. Associate Professor, Asia-Pacific School of Logistics, College of Business Administration, Inha University, 100 Inha-ro, Nam-gu, Incheon, , Republic of Korea. Phone) , Fax)

2 But, in existing literatures, financial efficiency is often measured by monetary values of inputs and outputs, and is used as an alternative method of a physical productivity measure. Though it is understood that these are partly due to the limitation of available data, these two methods may not show exactly the same results even with the same subjects of the same periods, because the different price structures among observations will affect the process of transforming physical values to monetary ones. Hence, the main aims of this paper are to figure out whether the physical productivity is consistent with the financial efficiency measured by the methods in existing literatures and, if inconsistent, to identify the origins of the inconsistency. The secondary aims are to evaluate the performances of the hub airports in the Asia-Pacific region with those two methods and to find out common trends of efficiency changes among them. For these purposes, this paper reviews existing literatures on efficiency measure of airports and strengthen the points at issue of this paper in section 2. Section 3 explains the methodology of efficiency measure and the data used in this paper. The results of empirical analysis are provided in section 4 with the implications of them. Section 5 concludes the paper with the major findings and limitations. 2. Literature Review Existing methods of efficiency measurement 1 can be classified into 4 broad categories. The partial factor productivity (PFP) is a simple non-parametric method which chooses an input variable and an output variable and measures the ratio of output over input. This method is very easy to measure with relatively simple data set, and is adoptable to any sets of input-output combination regardless of production or cost, whereas it can t grasp overall efficiency of an entity and is only to be comparable to other input-output sets with the same units of calculation. More generalized methods are the total factor productivity (TFP) and the stochastic frontier analysis (SFA). While the former is a typical non-parametric method of productivity comparing the weighted index of output with that of input, the latter is a parametric estimation method of a production/cost function, in which the efficiency is measured by the distance of the observation and the frontier. Concerning airport efficiency, Oum et al. (2003) and following yearly researches of the Airport Bench Marking Report are the representative contributions using TFP, and Choo & Oum (2013) is the recent result using SFA. These methods have the advantages of measuring overall efficiency and of comparability among firms. But serious drawbacks of them is that they require lots of data and complicated process of measurement, and that only one output variable is allowed in case of SFA. The most popular method for efficiency measurement is the Data Envelopment Analysis (DEA) in the field of air transport researches. Starting from the research of Gillen & Lall (1997), numerous papers have used the method and various modified models of it. The strength of this method is that it allows multi-outputs as well as multi-inputs, and that various kinds of efficiency can be measured along with the characteristics of variables chosen, whereas a serious shortcoming is that the measured efficiency score is only a relative term depending upon the variables used and the observation groups. 128

3 Table 1: Literatures on airport efficiency using DEA Paper Sample Input(s) Output(s) Gillen & Lall (1997) 21 US airports Bazargan & Vasigh (2003) Yoshida & Fujimoto (2004) Oum & Yu (2004) Lam et al. (2009) Chi-Lok & Zhang (2009) Chow & Fung (2012) Ahn & Min (2014) 45 US airports Japan airports world major airports major Asia- Pacific airports airports in China airports in China world major airports No. and length of runway No. of employees No. of gates, etc. Airport and terminal area No. of runways and gates Amount of operating-nonoperating expenses Length of runway No. of employees Terminal area Access cost No. of full-time equivalent employees Soft cost input index Labor cost Capital stock Soft cost Trade value Length of runway Terminal size Length of runway Terminal area Land area Length of runway Passenger terminal area Cargo terminal area No. of carriers Pounds of cargos Amount of aeronautic and non-aeronautic revenues Non-aeronautical revenues Table 1 shows the list of recent literatures on airport efficiency using DEA method. From it, we can find out that various input and output variables are used depending upon the possibility of data collection in each study. Most studies reviewed use the number of passengers, number of carriers, and the volumes of cargo as output variables, except Bazargan & Vasigh (2003) and Oum & Yu (2004) which add non-aeronautical revenues to output variables. Input variables are more diversified by authors. Many researches use the size of physical facilities, such as the number or the length of runways, terminal areas, and the number of gates, as input variables. Number of employees is frequently used as well. A serious problem in existing researches is that they used the limited numbers of variables especially in input variables. As Lam et al. (2009) pointed out correctly, the input variables should contain at least one component of each labor, capital, and intermediates to avoid the possible biases of the results of efficiency. Most of the above researches lack the variables of intermediates except for the cases of Oum & Yu (2004) and Lam et al. (2009). Some researches use only the size of physical facilities as input variables (Chi-Lok & Zhang, 2009; Chow & Fung, 2012; Ahn & Min, 2014). Analogous principle should be adopted in output variables that they must describe all the outcomes of airports which are operated to transport passenger and cargo and to run their derivative business. But, most of researches did not consider the output variable of nonaeronautical revenues, which occupies more than half of airport revenues and is getting 129

4 more important in the operation of airports, except for Bazargan & Vasigh (2003) and Oum & Yu (2004). More important issues to be considered is that physical productivity and the financial efficiency should be clearly classified. All of above researches measure physical productivity by using physical results as output variables. But, some of them use several variables of monetary term as input variables. The most extreme case is Lam et al. (2009), which uses only monetary terms of input variables. It may be inevitable to use the monetary-term variables as inputs or outputs, if the properties of them can t be converted into physical terms. Soft cost in input variables and non-aeronautical revenues in output variables are the cases. But others should be converted into physical terms if it is for physical productivity. 3. Methodology and Data The DEA model used in this paper is the super-efficiency slack-based model (super-sbm), which is a non-radial model allowing the efficiency score more than 1. Traditional DEA models, like CCR or BCC, are radial models which measure the efficiency score by the distance from origin. But these models miss the slacks among the DMUs in calculating the efficiency scores. The non-radial slack-based model can consider the slacks in measuring the efficiency score. Another shortcoming of the traditional model, including SBM, is that there are too many DMU s of which the efficiency score is 1. The method of measuring super-efficiency is to draw a new efficient frontier after deleting one of the efficient DMU s and to calculate the distance between deleted DMU and the new frontier. By doing this, we can calculate a new efficient score of the DMU s which was 1, but are not less than 1. There are numerous articles using the DEA-SBM for efficiency measurement and most of them explain and describe the mathematical model 2 of the methodology. Hence this paper just provides the formula of SBM and the illustrative figure for super-sbm. The formula of SBM is as Where = the efficiency score = the vector of input slack = the vector of output slack 130

5 = the input i of DMU k = the output r of DMU k = No. of input = the vector of the weight Figure 1: Illustration of the super-efficiency slack-based Model In Figure 1, by using SBM, the efficiency score of DMU A, B and C is 1, and the distance between C and D is a slack of D, whose efficiency score might be 1 if the score was measured with other radial methods like CCR or BCC. But, if we draw another frontier line by using A, C and D after deleting B, we can calculate the distance between B and P, which is the super-efficiency score of B, which is bigger than 1, while those of other A and C are still 1. The data used in this paper is a panel data for 11 major airports or airport groups 3 in Asia- Pacific for the period of 2001 and 2013, in total 115 observations. Airports are chosen based on the Top 20 Hub Airports Ranking announced by ACI in 2014 and the possibility of data collection. As shown in Table 2, some airports (group) have only relatively a short period of data due to the limitation of data collection. Most of the data were collected from the annual report, financial statement, and traffic report provided at the internet home page of each airport (group). The data of some characteristic variables were collected from the reports of ACI, CAPA, and the Airport Benchmarking Report published by the Air Transport Research Society (ATRS) every year. 131

6 Table 2: Summary of the sample airport (groups) City Country Airport codes Operating Body Period Incheon Korea ICN IIAC Tokyo Japan NRT NAA Kansai Japan KIX, ITM (Osaka) NKIAC Beijing China PEK BCIA Shanghai China PVG, SHA SAA Guangzhou China 5 airports including CAN GAA Hong Kong China HKG HIA Taipei Taiwan TPE TIA Kuala Lumpur Malaysia 21 airports including KUL MAHB Bangkok Thailand 6 airports including BKK, DMK AOT Singapore Singapore SIN CAG All the observations contain the number of employees, the area of passenger terminal, the area of cargo terminal, the soft cost, the labor cost, and the depreciation cost as input variables. Output variables collected are the number of passengers, cargo volume, the aeronautical revenue, and the non-aeronautical revenues. Among them, the input variables for the measurement of the physical productivity are the number of employees, the area of passenger terminal, the area of cargo terminal, and the soft cost, whereas those for the financial efficiency are the labor cost, the depreciation cost, and the soft cost. The output variables for physical productivity are the number of passengers, cargo volume, and the non-aeronautical revenues, while those for financial efficiency, the aeronautical revenue, and the non-aeronautical revenues. This means that this paper strictly classifies the differences of variables for two methods. The soft cost is overlapped for the input variables between the two methods, and the nonaeronautical revenues, for the output variables. This is inevitable because the property of those variables cannot be transformed to physical terms and it is very difficult to find out other proxies for them due to the data limitation of the industry. 4. Empirical results and their implications Figure 2 shows the average efficiency scores of recent 5 year by airports. It looks that the physical productivity and the financial efficiency of airports show little difference except for some airports like Kansai, Guangzhou, and Kuala Lumpur. These airports have the common characteristics that they are not one airport, but the airport groups including neighboring regional airports. But if we look into the airports which show high efficiency scores in both methods and which consist only of one airport more precisely, we can find out that they also have big differences in the scores over time between the two methods (Figure 3). 132

7 Figure 2: Average efficiency scores of airports (Recent 5 years) p.p f.e Figure 3: Differences of efficiencies by the variable types (selected airports) This means that the results of those two methods cannot be compared with the same yardstick. We can guess that this difference has its origin on the different price structure of both inputs and outputs by countries, though this paper has discounted all the monetary term variables by Purchasing Power Parity (PPP) to minimize the possible biases from exchange rate and price level. To identify more on the origin of the difference, this paper estimates 2 more efficiency scores by crossing the input variables and output variables, and measure the Spearman s rank correlation coefficient and Pearson s ordinary correlation coefficient among 4 efficiency scores. The results show that the two coefficients have little differences and that the efficiency scores with the same input variables have much higher correlations than others (Table 3). 133

8 From this, we can conclude that the price structure of input variables are more influential to the efficiency scores of the airports. Hence if we want to compare the efficiency of airports, we should use at least similar input variables no matter what the output variables are. Another practical findings are that, in terms of physical productivity, the airports which lead the market are Taipei, Incheon, Guangzhou and Beijing, whereas Kuala Lumpur, Bangkok and Tokyo relatively suffer from inefficiency. This implies that the airports in China and neighboring countries are more efficient than those in Southeast Asia in terms of physical productivity. It can be explained by the facts that the huge demand for air transport from China are offset by the supply of airports around China, while the airports depending upon the domestic market suffer from the relative demand shortage. On the financial perspective, Taipei, Kuala Lumpur, Beijing, Hong Kong, and Incheon are relatively more efficient, while Kansai and Guangzhou are less efficient. Compared with the results of physical productivity, these results of financial efficiency implies that the trends of efficiency among airports by two methods are relatively similar, except for some cases which compensates the productivity inefficiency by its price structure (Kuala Lumpur), and which are the second tier hub airports in their countries (Guangzhou and Kansai). Table 3: The correlation coefficients among efficiency scores Pearson PI FO FI - PO PI - PO FI - FO PI - FO FI - PO PI - PO FI - FO Spearman PI FO FI - PO PI - PO FI - FO PI - FO FI - PO PI - PO FI - FO Average trend of efficiencies in the market as a whole is shown in Figure 4. From it, we can identify that two efficiency scores are getting converged over time, especially after This is so in the analysis by countries, too. 134

9 1.200 Figure 4: Average trend of Efficiencies (super-sbm) P.P F.E 5. Conclusion This study has tried to figure out the consistency between physical productivity and financial efficiency, and to find out the origins of the difference. The results show that two methods are clearly different in their properties, and, should not be used for a mutual alternative. The difference comes mainly from the difference in input variables. Existing literatures often use the input variables of monetary terms in physical productivity measures, and compare the results of efficiency measure without considering the differences in input and output variables. But, the results of this paper show that they are not the proper ways. If we want to measure and to compare the efficiency of airports from different countries or over time, collection of input variables with the same or at least similar properties is utmost important and the physical productivity and the financial efficiency should be strictly classified. Despite those contributions of this paper, there is a possible way for making the results of the two methods have the same results. If the input variables in financial efficiency correctly reflect the price structure of each observation and the output variable is purely a total cost, the results of the two methods will be the same. But, in this case, the multiproduct characteristics of airports should be considered in different ways. These are left for future researches. Another finding of this paper is that the efficiencies of major airports in the Asia-Pacific region are getting better, either in productivity or in financial efficiency. But one important drawback is that the origins of the improvement cannot be identified by the analytical method of this paper. Regression analysis using the results of this paper as the dependent variable and other proxies on operation characteristics as the explanatory variables can be a useful method for the complementary research of this paper. This is left as a future research theme, too. 135

10 Endnotes 1. Oum et al. (2011) summarizes well the methods chronologically, especially adopting them to the efficiency measurement of airports. 2. Ahn & Min (2014) is a good sample reference. 3. The data named Bangkok is that of the Airport of Thailand Public Company Ltd. (AOT) which manages 6 major international airports in Thailand, and the data named Kuala Lumpur is that of the Malaysia Airports Holdings Berhad (MAHB) which manages all the airports in Malaysia. Kansai, Shanghai, and Guangzhou also contain the data of Osaka, Hongqiao, and neighboring regional airports, respectively. References Ahn, YH & Min, H 2014, Evaluating the multi-period operating efficiency of international airports using data envelopment analysis and the Malmquist productivity index, Journal of Air Transport Management, vol. 39, pp Bazargan, M & Vasigh, B 2003, Size versus efficiency: a case study of US commercial airports, Journal of Air Transport Management, vol. 9, pp Chi-Lok, AY & Zhang, A 2009, Effects of competition and policy changes on Chinese airport productivity: an empirical investigation, Journal of Air Transport Management, vol. 15, pp Choo, YY and Oum, TH 2013, Impact of low cost carrier services on efficiency of the major U.S. airports, Journal of Air Transport Management, vol. 33, pp Chow, CKW & Fung, MKY 2012, Estimating indices of airport productivity in Greater China, Journal of Air Transport Management, vol. 24, pp Gillen, D & Lall, A 1997, Developing measure of airport productivity and performance: an application of data envelopment analysis, Transportation Research Part E, vol. 33, pp Lam, SW, Low, JMW & Tang, LC 2009, Operational efficiencies across Asia Pacific airports, Transportation Research Part E, vol. 45, pp Oum, TH, Yamaguchi, G & Yoshida, Y 2011, Efficiency measurement theory and its application to airport benchmarking in A Palma et al. (ed.) A Handbook of Transport Economics, Edward Elgar, Northampton, pp Oum, TH & Yu, C 2004, Measuring airports operating efficiency: a summary of the 2003 ATRS global airport benchmarking research report, Transportation Research Part E, vol.40, pp Oum, TH, Yu, C & Xiaowen, F 2003, A comparative analysis of productivity performance of the world s major airports: summary report of the ATRS global airport benchmarking research report-2002, Journal of Air Transport Management, vol. 9, pp Yoshida, Y & Fujimoto, H 2004, Japanese airport benchmarking with the DEA and endogenous-weight TFP methods: testing the criticism of overinvestment in Japanese regional airports, Transportation Research Part E, vol. 40, pp