Energy Economics 25(2003)473–485
0140-9883/03/$-e front matter ᮊ2003Elvier B.V .All rights rerved.
PII:S0140-9883Ž03.00057-4A natural monopoly in natural gas transmission
D.V .Gordon*,K.Gunsch,C.V .Pawluk
Department of Economics,University of Calgary,2500University Dr.NW,Calgary,AB,
Canada T2N 1N4
Abstract
In this article,we test for subadditivity in the cost structure associated with transporting natural gas by Trans-Canada Pipelines Ltd.and measure for possible cost savings from incread competition that could be realized by removing the monopoly status granted by the National Energy Board.In measuring subadditivity,we apply both the Baumol et al.(Contestable Markets and the Theory of Industry Structure (1982))and the Evans and Heckman (Am.Econ.Rev.764(1984)613)procedures.Our results show evidence of subadditivity in the cost structure,and conquently,the possible benefits from incread co
学习通后台
mpetition resulting from splitting up the monopoly could be offt by the sacrifice of scale efficiencies.ᮊ2003Elvier B.V .All rights rerved.
JEL classifications:Q32;Q41
Keywords:Subadditivity;Natural monopoly;Natural gas transportation
1.Introduction
Over the past decades,numerous and significant changes have occurred within the natural gas industry,stemming largely from the deregulation of gas markets and pricing in both Canada and the US.In conjunction with enhanced competition in the marketplace,there comes an expectation to supply product to more distant markets as a means of supporting further growth in production capacity and profitability.As gas transmission costs reprent a significant portion of the total price of natural gas in many markets,it is not surprising that there is pressure to reduce gas transmission tolls as part of the move towards a more competitive commodity marketplace.
*Corresponding author.
E-mail address:dgordon@ucalgary.ca (D.V .Gordon ).
474 D.V.Gordon et al./Energy Economics25(2003)473–485
The incread focus on efficiency and cost issues with respect to gas has rved to focus the regulatory debate surrounding the pipeline industry.In particular, persistent and strident demands for further deregulation have been levied by the shipping industry,which contends that a shift in the regulatory regime would markedly improve efficiency and reduce tolls.The critical question underlying the issue is whether scope or scale economies would be sacrificed upon the advent of competition,thereby resulting in substantive inefficiencies that more than offt the benefits resulting from incread competition.An inquiry into this question requires knowledge of the cost characteristics in gas transmission and specifically whether natural monopoly is the natural structure of the natural gas industry. Subadditivity plays the dominant role in defining a natural monopoly.Subadditiv-ity is realized if no combination of multiple firms can collectively produce industry output at lower cost than a monopolist(Berg and Tschirhart,1988).In the single output ca,the existence of economies of scale is a sufficient condition to ensure subadditivity.However,in the multiple output ca the conditions for subadditivity are more elaborate requiring cost complementarities or product specific scale economies and economies of scope(Baumol,1977;Baumol et al.,1982).The interest by applied economists in subadditivity is in me
asuring the benefits from splitting up a monopoly industry into a number of smaller firms.If subadditivity prevails,no cost benefits will be realized by incread competition.On the other hand,without subadditivity(or what is referred to as superadditivity)lower costs of production could be realized by splitting up a monopoly industry into a number of smaller firms.
Much of the applied work in testing for natural monopoly has been carried out for the telephone industry(Smith and Corbo,1979;Fuss and Waverman,1981; Evans,1983;Evans and Heckman,1984;Roller,1992;Shin and Ying,1992).In fact,some of this work was instrumental in the decision to break up the Bell telephone system in US(Shepherd,1990).The natural gas pipeline industry is also an interesting candidate for natural monopoly testing but has received much less attention by applied economists(Aivazian et al.,1987;Ellig and Giberson,1993). In Canada,Trans-Canada Pipelines Ltd(TCPL)has been granted by the National Energy Board(NEB)of Canada an effective monopoly in the transportation of natural gas from fields in Alberta to points in Eastern Canada and the US.In spite of the relative paucity of empirical studies on subadditivity in natural gas transmis-sion,it ems reasonable,given the fundamental operating characteristics of the pipeline business,to speculate that subadditivity prevails in the cost structure of firms such as TCPL.In fact,Manll and Church(1995)argue that in natural gas
屠乃奔倚其下
1
transmission both plant and firm level subadditivity is prent.The purpo and contribution of this article is to asss components of the natural gas transmission industry in Canada by empirically testing for subadditivity in the cost structure
475
D.V .Gordon et al./Energy Economics 25(2003)473–485associated with transporting natural gas by the carrier TCPL and to determine whether cost savings could be realized by removing the monopoly status granted by the NEB.
In a multiple output tting,Baumol et al.(1982)and Evans and Heckman (1984)provide alternative procedures for testing for natural monopoly.The former approaches the problem indirectly by testing the production structure for the necessary and sufficient conditions that would indicate subadditivity.The latter is a direct test of the cost structure.In Section 2,the testing procedures are
建立信任
reviewed.In Section 3,we provide a brief overview of the Canadian natural gas pipeline industry,discuss the data available for analysis and specify the cost function ud in estimation.In Section 4,we evaluate the estimated cost function in terms of the regularity conditions and confirm that a ‘proper’cost function has been identified.Following this,the alternative tests for natural monopoly are carried out and results prented.Section 5concludes.
2.Methods
Subadditivity can be illustrated in a straightforward manner using a cost function reprentation of the technology.Let C (Y )reprent the cost function for a single 2firm existing in the industry and let C (y )reprent the cost function for the i th i firm in a multi-firm configuration of the industry.Subadditivity characterizes the cost structure if
C (Y )-C y (1)Ž.i 8where S y s Y .Changing the inequality to an equality implies additivity in the cost i structure and a further change to greater than implies superadditivity,in which ca,multiple firms can achieve lower total costs in producing industry output compared to the existence of only one firm.
Baumol et al.(1982)test for a form of local subadditivity.The procedure is to 3evaluate the estimated cost structure at each data point for evidence of subadditivity.In the single output ca,a test for econ
omies of scale is sufficient evidence to ensure a natural monopoly.In the multiple output ca,a sufficient condition is to find evidence of both economies of scale and transray convexity.形容情绪的词语
In the single output (y )ca,a measure of scale economies is the ratio of average to marginal cost (C )or
新西兰留学中介
y C (y )
S (y )s (2)
yC y where S (y ))1implies increasing returns to scale and the opposite inequality implies
476 D.V.Gordon et al./Energy Economics25(2003)473–485
Fig.1.Transray convexity.
decreasing returns to scale.Panzar(1989)extends this definition to the multiple
output ca or
C(Y)
S(Y)s(3) 8
y C(Y)
j j
where Y reprents an aggregate measure of total output and C(Y)reprents the
j
marginal cost of producing the j th output.Eq.(3)is defined for a ray from the origin where along the ray average costs are decreasing(Waterson,1988).Hence, if output-weighted marginal cost is below total cost,then S(Y))1and economies of scale are obrved.Converly,a less than inequality implies diconomies of scale.In empirical analysis,this scale measure is evaluated at each data point. Tran
sray convexity,in the multiple output ca,can be explained with reference to Fig.1(Evans and Heckman,1984).In the figure,ABCD reprents the cost
surface of a two output(y,y)firm with a ray from the origin labeled Db.Another
12
ray,a transray,intercts Db along abc.When the transray is extrapolated up to the cost hyperplane,it cross-ctionalizes the cost surface at ABC.A cost surface is said to be transray convex if the extrapolation of any transray up to the cost surface yields a convex cross-ctionalization.
Along the transray abc,let C(a)reprent a point on the cost hyperplane between (a and c)and let C(b)reprent a point on the cost hyperplane between(b and c). Transray convexity implies that a line connecting C(a)and C(b)must not lie below the cost surface at every point between a and b,or
C(k a q(1y k)b)(kC(a)q(1y k)C(b),0-k-1(4)
477
D.V .Gordon et al./Energy Economics 25(2003)473–485In other words,transray convexity implies tha
t holding the aggregate output bundle fixed but allowing a change in the output mix will lower costs for a diver rather than specialized output mix (Squires,1988).
Obrving transray convexity at one point in the data t does not imply convexity throughout the cost surface (Baumol et al.,1982).Squires (1988)suggests testing at each data point for pairwi transray convexity,where such pairwi evidence would imply that transray convexity holds overall.Pairwi transray convexity can be tested by calculating cond-order partial derivatives or
(a )Cy y 00,Cy y 00,Cy y s Cy y (0;
i i j j i j j i (5)
y (b )Cy y (0,Cy y (0,Cy y s Cy y (0,Cy y (y C y y Cy y c c j j i j j i i j i i j j
where Cy y is the cond-order partial of the cost function with respect to the i th i i output.Cy y (0implies weak cost complementarities and Cy y (0reprents i j i i product specific scale economies.From the last argument in Eq.(5),product specific scale economies are consistent with transray convexity only if cost complementarity is sufficiently strong (i.e.Cy y (y 6(Cy y Cy y )).
i j i i j j It is interesting to note that cost complementarity reprents an alternative sufficient condition
for subadditivity and,what is more,implies economies of scope,which also reprents a necessary condition for subadditivity.Empirically,the 4Squires procedure involves testing both for transray convexity and for cost complementarities.国学典故
嘴唇薄的男人性格特点Evans and Heckman (1984)suggest a direct test for subadditivity.Here subad-ditivity is defined for the two output ca (y and y )with n firms in the alternative 12configuration of the monopoly industry as
C y ,y -C a y ,b y ,a 00,b 00,i s 1,«n
(6)
Ž.Ž.12i i i 2i i 8i If 8s 8s 1,the combined output of the individual firms equals the output of a b i i the monopolist (i.e.additivity ).On the other hand,if the inequality in Eq.(6)is reverd,the cost function is characterized by superadditivity.
In measuring for subadditivity,it is necessary to compare the actual cost structure of the existing monopolist with the cost structure that would apply to an n -firm configuration of the industry.To accomplish this,Evans and Heckman place restrictions on the admissible output vectors ud in testi
ng.In general,they impo the restriction that the n -firm output vectors are within the range of output vectors actually obrved in the data.Operationally,this implies that no firm in the alternative configuration of the industry is permitted to produce less of each output
478 D.V .Gordon et al./Energy Economics 25(2003)473–485than the lowest obrved output level of the monopolist.This impos the restriction that testing can only be performed in a region where the output level of each product is at least twice that of the minimum obrved production level of the monopolist.In addition,Evans and Heckman place a further restriction on the output mix for the n -firm configuration to be within a ratio obrved in the data.This restricts any individual firm from having a more specialized output mix than the monopolist.
Imposing the Evans–Heckman restrictions,we rewrite Eq.(6)as
n
i R R C y ,y y C y ,y Ž.Ž.12128i s 1
Sub s (7)
C y ,y Ž.12where C reprents the i th firm’s cost function and y is the j th output greater than i R j or equal to the minimum level of this output obrved in the industry.If Sub -0,the data suggest subadditivity in the cost structure and,on the other hand,if Sub )0,the data suggest superadditivity.
3.Natural gas transmission
The NEB of Canada regulates the transmission of natural gas within Canada with respect to tolls,expansions,allowed rate of return and the right to transport.It is the latter certification that can bestow monopoly status in the natural gas transmission market.TCPL falls within the jurisdiction of the NEB and has been granted a certificate to transport natural gas from fields in Southern Alberta to delivery points in Canada and in the US.TCPL’s main transmission line started operations in 1958.Prior to 1986,TCPL was engaged in both transporting and marketing natural gas.In 1986,the NEB deregulated the pricing of natural gas and withdrew the right of pipeline companies to engage in marketing and sales of natural gas.In 1986,TCPL formed Western Gas Marketing to carry out the marketing and sales of natural gas.Currently,TCPL has 13687km of pipe delivering some 55.8billion cubic meters of natural gas annually.
呼叫英文
Conditions within the natural gas transmission industry appear to reflect plant level subadditivity,or subadditivity that is associated with cost-efficiencies due to indivisibilities in production technology resulting in economies of scale and scope (Manll and Church,1995).In the pipeline industry,volumetric returns to scale 5exist such that as the diameter of the pipe doubles,its volume increas by a factor of 4,while its surface area only increas by a factor of 2.Output is proportional to volume;however,the cost of construction is proportional to surface area.Economies of scope can ari if there are large common costs in the production of
