Kobe University Repository : Kernel

Transcription

1 Kobe University Repository : Kernel タイトル Title 著者 Author(s) 掲載誌 巻号 ページ Citation 刊行日 Issue date 資源タイプ Resource Type 版区分 Resource Version 権利 Rights DOI URL The Quality Competition among Japanese Major Airlines after the Regulatory Change of 1985 Murakami, Hideki The Annals of the School of Business Administration, Kobe University,40: Departmental Bulletin Paper / 紀要論文 publisher Create Date:

2 The Quality Competition among Japanese Major Airlines after the Regulatory Change of 1985 Hideki Murakami * Abstract This paper analyses econometrically the mutual quality competition of three major Japanese airlines (JAL, ANA, and JAS), after regulatory change of Using 31 cross section data of double and triple track markets for 4 years, the main section reveals that (a) through quality competition JAL has successfully increased its share of demand in the markets it was authorized to enter, while ANA, the incumbent in those markets, has gradually lost its share of demand, and (b) with respect to JAS, it has also been able to increase its share of demand, but less successfully than JAL. The final section refers to the character of the aviation policy of the government which may have caused the different performance in the quality competition between these airlines, and investigates why such a difference arose. * Assistant Professor of School of Business Administration, Kobe University. The author is thankful to Professors Tae Oum (University of British Columbia), Anming Zhang Brunswick),and (University of Victoria), Chin Anthony (National Yiming University Zhang (University of New of Singapore), who gave me some helpful comments at the 7th World Conference on Transportation Research in Sydney. A part of the contents of this paper is repeated in Morrison et al.(1995).

3 62 H. Murakami I. Introduction It was in 1972 that the Japanese government (Ministry of Transport, the MoT) decided on the aviation policy, called the system. The main idea of the system was that JAL (Japan Air Lines, owned by the government) was authorized to operate mainly in international and domestic trunk markets, ANA (All Nippon Airways, privately owned) mainly in domestic trunk and the denser local markets, and JAS (Japan Air System, privately owned, which once was called TDA) mainly in other domestic local markets. Under the system, the MoT also tightly regulated all kinds of competition, such as price competition, entry and exit, and quality competition. However, in 1985, influenced by the world-wide movement toward deregulation, the MoT revised the system. In particular, with respect to domestic policy changes, the MoT authorized JAL and JAS to enter some dense domestic markets where ANA had already been operating as the incumbent, so that 34 double and triple track markets were formed by the end of 1992'. However, it is clear that the MoT continues to regulate the competition mentioned above, even after the regulatory change. So until now there seems to have been no room for airlines to compete with one another in domestic markets. However, it is interesting and a little strange that the MoT declares that it intends to accelerate, at least, the growth of the spirit of competition through regulatory change. So the MoT's purpose in its regulatory change looks obscure, or from another viewpoint, looks complex. Before conducting a theoretical and empirical analysis, it is necessary to explain the structures of this "strange" aviation policy, which is done in the 1. Trunk markets are those which connect Sapporo, Tokyo (Narita, Haneda), Osaka (Itami, Kansai International), Fukuoka, and Naha to one another. Among the domestic markets (i.e., excluding the markets that serve Narita and Kansai International), there are nine trunk markets in " double (triple) track" markets are the routes where two (three) airlines operates simultaneously. Before the regulatory change (i.e., under the system), in order for a monopoly market to be double track, more than 700,000 passengers per year were required, and for a double track market to be triple track, more than 1,000,000passengers per year were required in principle (there were some exceptions). However, after the regulatory change, the MoT revised the criteria, so that for, a monopoly (double track) market to be double (triple) track, the threshold was lowered to 400,000 (700,000) passengers per year. In accordance with the constant increase in the number of passengers after the regulatory change, double and triple track markets have been formed more rapidly than under the system.

4 The Quality Competition among Japanese Major Airlines after the Regulatory Change of following section II. Taking into consideration the context of section II, section III tries to construct the "mutual" quality competition models. Section IV shows the empirical results of the models constructed in section DI, where it is revealed that each airline mutually competes in the different manner under the circumstance it faces. The final section refers to the aviation policy of the MoT that may have caused the different type of quality competition between the three airlines, and investigates why such difference arose. II. The structures of quality competition With respect even to the quality competition, Japanese airlines are forbidden from competing directly in domestic markets. However, the MoT intended to let airlines compete with each other, especially after the regulatory change. So it seems that what the MoT really intended was to accelerate the development of the indirect or mutual quality competition among/between airlines. The simple concept of mutual competition is shown in Figure 1. mutually compete airline A (1) "airline A" negotiates with the MoT, expecting the government to authorize it to operate in some profitable markets. negotiaty Government authorize, (MoT) airline 13 (2) Considering the vested interests of both the airlines, the MoT authorizes "airline A" to enter profitable markets, and at the same time reassures "airline B" and then promises to give B another profitable market in the next stage of the negotiation. Figure 1 The simple concept of mutual competition between two airlines

5 64 H. Murakami The essence of Figure 1 is that airlines compete not directly with each other but indirectly with the MoT intervening between them. So the MoT has to continue to prepare the "next stage" where it reassures one airline by offering compensation for authorizing the other to enter new markets, so long as airlines continue to take turns applying for authorization to operate in profitable markets. Strictly speaking, regarding the quality competition in the double or triple track markets, what airlines aim for in the long run through the negotiation with the MoT is how and to what extent they can increase their number of their own departures at the peak demand hours in a day. However, there is a significant problem with this ; the capacity limitation in the major airports (e.g., Haneda and Itami) prevents airlines from increasing the number of departures. So, in order to increase the amount of their own supply, they need to expand the aircraft size ; for example, they have introduced such wide-bodied aircraft as the B747 class, DC-10 class, or L-1011 class (in the case of ANA) at the peak hours in a day. And they have also tried to change flexibly the type of aircraft according to the change in demand in a day, month, and year. Summing up the character of the quality competition, airlines compete with one another for the purpose of (a) increasing the number of their own departures and (b) introducing wide-bodied aircraft (best fitted for the demand), at the peak demand hours in a day and, sometimes, on the peak demand days in a month, and in the peak demand months in a year. M. The models of quality competition and the data In constructing the models of quality competition inherent in the Japanese domestic airline industry, this paper refers to the quality competition model introduced in Douglas and Miller (1974b), which modeled and tested the effect of the quality competition before deregulation on the producer's surplus'. In the case of the quality competition in the United States, airlines tried to attract passengers by increasing only their share of departures in a market. However, as is mentioned above, unlike the US airline industry before deregulation, 3. See Douglas and Millerlif(1974b), pp and Schmalensee (1977), pp

6 The Quality Competition among Japanese Major Airlines after the Regulatory Change of Japanese airlines have to try not only to increase both the number of departures but also to increase the number of seats per flight, in order to win the quality competition. Taking the character of this competition into account, this section constructs models that are unique to Japanese airline industry. Suppose that the total cost of an airline i (TC ) is simply dependent on the number of passengers carried by i (PAX ) and the total output, i.e., available seat kilometers (ASK i )- TCi==f (PAXi,ASKi) To make the problem more simple, we suppose a linear relation between total cost and the independent variables, i.e., ( 1 ) * PAXi+ q * ASKi where p * PAXi represents the cost of serving the passengers, and q* ASKi is the cost of operating aircraft. Next we suppose the following demand function. ( 2 ) PAXi= f (F,INC,POP,V,),where PAL is the demand carried by airline i, F is the regulated fare, INC is the average income of the passengers carried, POP is the market size (population) of a market, and V i the service vector of airline i. To simplify the model, the service vector here is defined as the total number of seats (the number of departures multiplied by the average number of seats per flight) served by an airline i (S 1)4. So the profit function of firm i is as follows. ( 3 ) ri=f*paxi TC,=-F* PAXi (p *PAXi+q * ASKi) Then suppose airlines maximize their profit with regard to the total number of seats (Si), the first order condition of equation (3) goes as follows' : 07r ( 4 ) 0iasi SiaPAXi (F p) q * DIST= 0 4. For example, the more the total number of seats increases, the easier it is for passengers to make reservations. 5. Note that available seat kilos are the product of the stage length and the total number of seats, soa ASK,= as DIST.

7 66 H. Murakami,where DIST is the stage length of a market. And the regulated fare under airline i's profit maximization is described as : *- ( 5 ) F -qdist a PAX+p a s. On the other hand, if airline i charges the fares at the break-even level, i.e., 0, then we can rewrite equation (3) as: ( 6 ) F = q * DIST +p PAX, Si Then we can obtain the condition from equation (5) and (6) that airline i earns profit under the regulated fare: a PAX, 0 S. PAX Next, we consider the share of passengers carried by an airline. S,s,s. s,= s, (i = 1, 2, n) E Sk E Sk E Sk k=1 k=1 k=1 where "n" is the number of airlines in a market. So, provided that the load facter is equal among the airlines, the relation between the number of passengers carried by airline i and the total demand in a market is described as : PAXi= s,* E PAX,, k=1 When airline i intends to maximize the number of its passengers by adjusting its total number of seats, i.e., differentiating PAX i with regard to Si, we can obtain : a PAX.k=1PAXk as; 7 )( *ae+ PAXk * as;as; k=1 as. S apaxk ae.k=1k= 1 1 S E PAXk I * ih=1sk sh ask asp k=1 E S,as, S k=1 k=1 k=1k k=1 az 12

8 The Quality Competition among Japanese Major Airlines after the Regulatory Change of s. aesa k= 1 E AaSkSi k=1 A ae PAXk k=1 * A ae sk It- 1 A si 1 ESk k=1 az s k=1 asi * E PAXk k=1 E Sk k= 1 a E P AXk k=1 a i a k=1 To simplify equation (7), we describe ki issk-s,li and ob- a SPAXk-=-7-PAX, viously=1, so equation (7) can be rewritten as : as ; a PAXiIS.+(iSi ) *[PAX I apax1}()pax s, PAX la PAX ( 7 )' a s,sssa sa s If we describe =---s SY X,and= a s in equation (7)', the terms in parenthesis { } are : X+ ( 1 X) *Y In the double or triple track markets, it is obvious that 0 <X< 1 0 PAX PAX, and in addition, if profit exists for airline i, i.e., a s i s, Y> 1 can be obtained. In this case, the term {X+ (1 X) *Y} is always larger than unity. With reference to Douglas and Miller (1974b), this suggests that an additional increase in airline i's share of the total number of seats in a market leads to a greater increase in its share of passengers. And if each airline recognizes this phenomenon, it is suggested that it will continue to increase the total number of seats until the point of zero rent " is reached, which will lead to the low load factor (i.e., the waste of resources)6. 6. Being different from the definition in this paper, the quality competition in Douglas and Miller (1974b) is defined as the "frequency competition". Panzar(1979) also refers to this type of quality competition. See Panzar(1979), pp

9 1 68 H. Murakami In the econometric model, this is described as the following loglinear function : PAX I=S ( 8 ) LnPAX= a+b* Ln + where "a" is the constant, b X+ (1 X)* Y> 1, and ft is the error term. Now notice that the total number of seats is the product of the number of the departures and the number of the seats per flight. And remember that the Japanese airlines can adopt only two ways; one is to gain the departure slots and/or the other is to expand the aircraft size at the peak demand hours in a day in order to win the quality competition. Then suppose airline i's passenger share function is : {FRQi I 7 *r (9) si = a FRQ EQ where FRQ, /FRQ is the share of departures at the peak demand hours, and EQ /EQ is the share of the average aircraft size at the peak demand hours that may represent the attractiveness (or strength) of i's service quality. In this equation, Japanese airlines are supposed to make the best of (i.e., optimally allocate) their managerial resources under the circumstances they face, in other words, they are supposed optimally to choose to increase either their own share of departures or their share of the average aircraft size at the peak demand hours, or to increase both of these. Following the context of this model, we suggest that b= r + q5, and if a profit exists for airline i, then b= r + cb> 1. In this case, an airline can be regarded as being successful in attracting passengers by making the best of its managerial resources. And for example, if is larger than r and both of them are statistically significant, we judge that the airline emphasizes the strategy of upgrading or expanding its fleet at the peak demand hours. This phenomenon is to be expected if the markets used as the samples are seriously congested, and in particular may also be true of ANA which has been the incumbent and is no longer able to increase the number of its departures. Converting the equation (9) to the econometric model, it can be rewritten as :

10 The Quality Competition among Japanese Major Airlines after the Regulatory Change of Ln a kADh +r +(LAD] PAX) Ln FAX = izfrq [ iLAD]Ln FEQQ' + k9e where ADk is the year dummy variables (k =89, 90, 91, and 1992 is the benchmark). The rest of this section introduces the data used in the following analysis. The markets selected here were both domestic double and triple track markets at the end of 1992 (see Table 1). Table 1. The domestic double and triple track markets, the firms in operation for the period '89-'92. and the number of City Pair Route Firms City Pair Route Firms 1. Tokyo/Sapporo Tokyo/Nagasaki Tokyo/Osaka Tokyo/Kumamoto Tokyo/Fukuoka Tokyo/Oita ANA,JAS 4. Tokyo/Naha Tokyo/Miyazaki ANA,JAS 5. Osaka/Sapporo Tokyo/Kagoshima Osaka/Fukuoka JAL,ANA 22. Osaka/Sendai JAL,ANA 7. Osaka/Naha JAL,ANA 23. Osaka/Matsuyama JAL,ANA 8. Fukuoka/Sapporo JAL,ANA 24. Osaka/Kochi ANA&ANK,JAS 9. Fukuoka/Naha JAL,ANA 25. Osaka/Kagoshima Tokyo/Kushiro ANA,JAS 26. Nagoya/Sapporo JAL,ANA 11. Tokyo/Hakodate JAL,ANA 27. Nagoya/Sendai ANA,JAS 12. Tokyo/Akita JAL,ANA 28. Nagoya/Fukuoka JAL,ANA 13. Tokyo/Kanazawa Nagoya/Nagasaki ANA,JAS 14. Tokyo/Hiroshima Nagoya/Kagoshima JAL,ANA 15. Tokyo/Takamatsu ANA,JAS 31. Nagoya/Naha ANA,JAL&JTA 16. Tokyo/Matsuyama JAL,ANA Note: (1)The markets where ANA and ANK or JAL and JTA operates are not regarded as double track, because they are horizontally integrated. (2)An airline which operates for less than 6 months in a year is not included in each annual datum. (3)JAL,ANA,and JAS operate in all the triple track markets. (4)The following local markets are excluded because of their smaller market size; Fukuoka/Kagoshima, Kagoshima/Amami, and Naha/Ishigaki The data comprise thirty one cross section data for four years. The remarkable character of the double and triple track markets is that (a)

11 70 H. Murakami ANA is always the "first corner", except for the Fukuoka/Sapporo market in 1989 and 1990 and it has the large share especially in triple track markets except for trunk markets (see Table 2). (b) JAL is the "first corner" in the trunk markets, i.e., No.1-9 in Table 1), and the "second corner" in the oth er local markets it was allowed to enter after the regulatory change, and (c) JAS is sometimes the "first" or the "second" corner in some local markets, but always the "third corner" in the trunk markets, and its ratio of the share of those markets is considerably smaller than that of the other two airlines (the average shares of JAL, ANA, and JAS is about 4.5 : 4.5 : 1, respectively, see Table 2). Table 2. The major airlines' output share (% ) in each triple track market in 1992 City-pair-routes JAL ANA JAS City-pair-routes JAL ANA JAS Taneda/Shin-Chitose aneda/itami Haneda/Fukuoka Haned a/naha Itami/Shin-Chitose Haneda/Komatsu Haneda/Hiroshima Haneda/Nagasaki Haneda/Kumamoto Haneda/Kagoshima Itami/Kagoshima Note : calculated by the author. The data source is Koku Yuso Tokei Nempo So the number of samples of each airline is not always 124; the number of sample of JAL is 76, ANA 122, and JAS 68, respectively. Each datum is not daily or monthly aggregated but annually. However the effect of the competition to win the peak demand hours in a day is expected to appear even though we use annual data. The data sources are Koku Yuso Tokei Nempo, , the handbook filing the annual operational data of all the domestic markets, published by the MoT. N. The empirical results The empirical results of equation (9) are shown in Table 3 on the next page. All the equations are estimated by OLS, and the ones in the upper row of each airlines' equation are estimated without any year

12 The Quality Competition among Japanese Major Airlines after the Regulatory Change of Table 3. The regression results of equation (9) a+ 1, r+ sk Ak adj. R2 SE ( 1) JAL (12.21a) (2.63b) (8.58a) ( 2 ) AD AD AD. D AD9, (8.61a)(3.51a) (12.63a)(3.49a) (2.74a) (2.16b)(2.04b) AD AD9, (2.80a) (2.05b) (1) ANA (5.85a) (0.28) (16.31a) ( 2 ) AD AD AD. (15.47a)(1.84) (5.16a)(1.43) (0.89)(1.56) AD. (1.32) ( 1) JAS (7.60a) (5.62a) (12.24a) ( 2 ) AD AD (7.95a) (6.50a)(2.73a) (9.19a)(2.66a) Note : t-statistics are in parentheses ; the level of significance : a=1%, b=5%. dummy variables. The estimation method of each equation in the lower row is, to begin with, to introduce all the year dummy variables both in the constant and in the coefficients and then, in order to improve the statistics, to remove such dummy variables as are not significant even at the 20% level. The result for JAL shows us that (a) JAL generally prefers to choose the strategy of introducing wide-bodied aircraft at the peak demand hours in a day, but the strategy it considers as important changes a little in each year, and (b) it has been successful in increasing its share of passengers through negotiation with the MoT, for b= r + 0>i. The reason why result (a) is obtained is probably that it is the incumbent in the trunk markets most of which are congested, so it can only choose to introduce the larger type of aircraft in order to increase its share of the passengers in these markets. However, it seems that it was also able to choose the strategy of increasing its share of departures when he was authorized to enter the local but dense

13 72 H. Murakami markets, which are less congested than the trunk markets, for the coefficient r is relatively larger in these years. With regard to ANA which already occupied the peak demand hours of the trunk and dense local markets, it is natural that there is no room for him to increase its frequency of departures. In order to attract passengers or to meet the increase in demand, it can choose only the strategy of introducing or upgrading the aircraft which are suitable for the demand in the markets where it already operates. By doing so, ANA has increased its share of market demand, but less successfully than the other two firms, for the value of 0 is closer to unity. The result for JAS is similar to that of JAL, but the value of the sum of the parameters, i.e., r + q is smaller than that of JAL. Indeed JAS was authorized to enter some of the trunk markets, but it has not been given the right to operate at the times when it wishes, because of the lack of slots at Haneda or Itami airports. In addition, the size of the aircraft JAS operates is a little smaller than the other two airlines. For these two reasons, JAS has not increased its share of the passengers as successfully as JAL ; in other words, JAS has not benefited so much from the aviation policy of the MoT or the regulatory change'. Finally, judging by these results, it is clear that the types of the mutual quality competition of Japanese airlines are different from one another. The results of the Chow test among the airlines are : JAL/ANA : F =20.15> F (3,192:0.00 = 3.88 JAL/JAS : F =6.66> F (3,138:0 01) = 3.93 ANA/JAS : F =8.29> F (3, ) =3.89 These three results show that the null hypotheses that each airline's type of mutual quality competition is equal are rejected at the 1% level. V. Concluding remarks As is mentioned in Chapter II, Japanese Ministry of Transport has adjusted the vested interests and flying rights of the airlines, and summing up the empirical results in section IV, this has led to different 7. And corresponding to the change of the demand, i.e., to the boom or slump, JAS is sometimes made to suspend the operation in some trunk or dense local markets (e.g., Itami/Fukuoka or Haneda/Hakodate).

14 The Quality Competition among Japanese Major Airlines after the Regulatory Change of performances in the mutual quality competition between the airlines. The airline that benefited most from aviation policy including the regulatory change of 1985 is JAL, on the ground that it has been able to increase its share of passengers through mutual quality competition (i.e., through negotiation with the MoT). ANA, which was the incumbent in all the trunk and dense local markets, has gradually lost its share of passengers. With regard to the empirical results of this paper, the authorization of the entry of JAL and JAS to the new markets has not benefited ANA. However, to compensate for these losses, the MoT has given ANA several long haul domestic monopoly markets, for example, between Sapporo (Shin-Chitose) and the western part of Japan, and also several long haul international markets, to and from North America and Europe. So, if we think of the overall effects of the regulatory change on the performance of ANA, it may also have benefited from the aviation policy of the MoT. With regard to JAS, it is difficult to judge whether it has benefited from the aviation policy including the regulatory change, because the MoT has not greatly changed its policy toward JAS even since the regulatory change; it was allowed to enter some trunk markets, but it has not been very successful in increasing its share of passengers through negotiation with the MoT. In addition, with regard even to JAS's entry into international markets, by which the MoT meant to compensate for its unsuccessful performance in domestic trunk markets, JAS has been authorized to enter only three markets that are shorter than those of ANA (i.e., Narita/Honolulu, Narita/Seoul, and Narita/Singapore). Further more, JAS still has many sparse markets even though it transferred some of them to its subsidiary, Japan Air Commuter. Judging by the treatment of JAS after the regulatory change, it can be said that the purpose of the regulatory change by the MoT has been to alter the vested interests of JAL and ANA in particular, but not of all three airlines. Finally, from the viewpoint of welfare analysis, the problem of excess capacity (i.e. a low load factor) as a result of the quality competition (mentioned in section III) will not arise, at least in the near future, unlike the US case before the deregulation of 1978, because of the capacity limitation in the major airports. However, for example, if we focus on the case of JAL, we have to pay attention to whether JAL cross-subsidizes the international competitive markets with the

15 74 H. Murakami increased revenue earned in the domestic markets it was authorized to enter. If this were true, the allocation of resources would be regarded as being distorted. It is suggested that what the MoT should intend is not the alternation of the vested interests of the airlines in an ad-hoc manner, but the prevention of the resources from being wasted. Received August 29, References Douglas, G.W., and Miller, J.C., Economic Regulation of Domestic Air Transport, Brookings Institution, 1974a. Douglas, G.W., and Miller, 111, J.C., "Quality Competition, Industry Equilibrium, and Efficiency in the Price-Constrained Airline Market", American Economic Review, Vol.64, No.4., 1974b. Morrison, S. A., Yarrow, G., Lawton-Smith, H., Yamauchi, H., and Murakami, H., "International Comparison of Privatization and Deregulation among the U.S.A., the U.K., and Japan Vol.1II Airline and Trucking", Keizai Bunseki No.143, Economic Planning Agency, Murakami, H., "The Effect of the 1985 Regulatory Change on Quality Competition and Cost Structure in Japan's Airline Industry", Kobe University Discussion Paper Series 9517 (submitted to 7th World Conference on Transportation Research), Panzar, J.C.,"Equilibrium and Welfare in Unregulated Airline Markets, American Economic Review, Vol.69, No.2, Schmalensee,R.,"Comparative Static Properties of Regulated Airline Oligopolies", Bell Journal of Economics, Autumn, 1977.