Special Interest Group on Maritime Transport and Ports
a member of the WCTR Society

INTERNATIONAL WORKSHOP
Genoa - June 8-10, 2000

Indice Relazioni

Dedicated containers terminals:
costs and benefits from a port perspective

Drs Benacchio - University of Genova Drs Cariou - University of Nantes Professor Haralambides - Erasmus University of Rotterdam

One of the main trends during the last ten years is the emergence of container dedicated terminal in ports. This paper tries to analyse the reasons that could explain this new trend and the implications from the port point of view. A first section presents an overview of the dedicated terminals in the world and underlines that a general definition of a dedicated terminal is difficult to find. A second section stresses through the use of a generalised port cost function that one of the main factor that could explain the emergence of dedicated terminals can be the increasing gap between port and shipping lines objectives. Finally, a third section analyses, through the use of a simple queuing model, under which assumptions a dedicated terminal could be a "win-win" strategy for both port and shipping lines.

Introduction

The move to dedicated terminal is not a new one in port activities. In bulk shipping for example, the need for special infrastructures, the safety regulation… have already implied the segmentation of port area by activities between liquid terminals, bulk terminals…

The development of containerisation and its impact on the organisation of ports during the sixties also leads to a spatial differentiation between container terminal area, and others terminals ([1], [2]). The emergence of dedicated terminals is not therefore referring to the specialisation of terminals that already exists, but rather to a new trend that leads to the discrimination within a container terminal between users. The emergence of dedicated terminals arises in a period of horizontal consolidation between ship-owners and of vertical agreements ([3], [4], [5], [6], [7]). This evolution is mainly a consequence of the requirement from shippers of a higher geographic coverage and a better "supply chain management" ([8], [9], [10], [11]).

1. The definition and emergence of dedicated terminals

A dedicated terminal can refer to a geographic location, the use of facilities for a defined part of a terminal, to a temporal exclusivity, the use of facilities for a certain period, and usually both. The scope of a dedicated terminal can for example implied the exclusivity for a shipping line of quay, berth, stacking area, railway connections… within port and is difficult to generalise. In a sense, the use of "windows" for containerships already represents a time-limited dedicated terminal.

Scope of dedicated terminals

An other difficulty with the definition of a dedicated terminal arises from the definition of the content of the contract. For example, what are the equipment (cranes, straddle carriers…) and for which duration, the contract takes place?

A unique definition of a dedicated terminal in container industry is therefore difficult to reach. It implies a private contractual commitment between the port operators that could be a stevedoring company and/or a port authority, and the port users i.e. the shipping lines.

This private contract can lead the Liner Company to do the stevedoring by itself (see for example Maersk in Algeciras and Rotterdam) or to let a stevedoring company to run the terminal (MSC and CP Ships with Hessenatie in Antwerp for example).

A way to analyse the emergence of dedicated terminals is through the analyse of their locations. Figure 1 presents some examples of dedicated terminals in Asia, North America and Europe in 1998.

Figure 1: Main dedicated terminals in Northern America, Asia and Europe

Source: Bank of Japan, Containerisation International, Lloyd's List

The main common element when facing the emergence of dedicated terminals is their strategic location. Two components have to be considered: the strategic geographic location on the main trades or in interconnecting points between East-West and North Trade; the strategic location in the network of the shipowners.

It appears that to be one of the biggest ports in the world in terms of throughput is an important but not a sufficient condition for the development of a dedicated terminal. Evergreen or Maersk for example initially decided to develop a dedicated terminal in Sines or in Algesiras not because they were the biggest ports in the world, but because they were to become one of the most important ports in the organisation of their own network.

The next section discusses some of the reasons that could explain the emergence of dedicated terminals and underlines some implications from a port point of view.

2. Cost and benefit from a port point of view

A way to analyse the demand by shipowners of dedicated terminals can be through the use of the generalised port cost function. The generalised port cost function integrates the value of time for users and shows the gap between the level of traffic that minimises the cost for port (q₀₂) and for the port and the shipowners (q₀₁).

Due to the high level of fixed cost for the port, as long as the throughput is increasing, the cost per ton is going to decrease till a certain point where the congestion costs imply an increase in the average cost. For the ship, the value of time in port is going to increase with the size of the vessel and the kind of cargoes it carries [11].

Figure2: Effect of the increase in vessel size

A direct consequence of the growth in vessel size that has occurred during the last ten years ([12], [13]) can be seen in the increase of the value of ship time cost in port (from Ship time cost 0 to 1 in the figure). This increase induces a wider gap between port and shipowners objectives (q₀₂ - q₁₁ > q₀₂ - q₀₁).

From this relation, it can be stressed that for the shipowners who aim at minimising the ship time cost, the use of a dedicated terminal can be a way to control the previously exogenous factors that play on the time in port for the ships. The dedicated terminal gives the shipowners the opportunity to reduce the expected impact of queuing time in port, to minimise the impact of a delay in the arrival of its vessels, to plan in a more efficient way its schedule, and notably to co-ordinate mother and feeder vessels. The willingness for shipowners to "secured" space and time within a terminal can not solely be motivated by physical aspects. The fear for example of information sharing and the need to protect confidentiality could also lead to a demand for dedicated terminals.

The development of dedicated terminals is therefore generally present as the consequence of a higher shipowners' pressure on port.

At the same time, the emergence of dedicated terminals can also be seen as an attractive factor of competition among smaller ports. Evergreen chose for example Sines (Portugal) and Tarento (Italy) because of the possibility to create dedicated terminals in those ports rather than in ports close to full capacity.

It can be also argued that, over a certain level of throughput q₀₂, the cost per ton is going to increase. Under the assumptions of a level of traffic higher than this breakeven point and of fixed capacity in the short run, the reduction in occupancy rate does not automatically imply a loss for a port.

Furthermore, the pressure on ports implied by the consolidation of shipowners can not be analysed as a pure exogenous factor for a port. Its ability to undertake for example new investments, new railway connections or inland terminals can reduce the gap between shipowners and port' operators objectives. Therefore, the need of shipowners to secure space and time can also be fulfilled by port operators' actions.

Finally, it appears that the stevedoring companies and the port authorities also play an important role in the emergence of dedicated terminals. They arose in a general context of worldwide investments in new capacities and disengagement of public funding (port authorities, regional, state). In order to secure investments, the new concession tenders take into account the possibility from potential operators to bring goodwill (see for example the concession in Antwerp for MSC/Hessenatie, CP Ships/Hessenatie) and are also encouraging stevedoring companies to choose for dedicated terminals.

3. The dedicated terminal: conflicting or common interest?

A way to show the trade off between shipowners' and port operator' objectives in the emergence of dedicated terminal is through the use of the Queuing Theory ([14], [15]).

In a general queuing theory system, the occupation rate is defined according to the probability of arrival rate λ and service times μ (usually, a Poisson distribution). In the simplest case (one stage, infinite queue and First In First Out), we can consider the choice for a port to keep a multi-user terminal with m servers, or to split the terminal with (d) dedicated servers and (m-d) multi-user servers.

From the port's point of view, the effect of moving from the first situation (pure multi-user) to the second one (multi-user and dedicated) is equivalent to comparing the occupation rate for the first (₁) and the second (₂) case with:


(1)

If we assume that the service rate is not going to change with the numbers of servers (constant economies of scale on services rates), and if the split of ships follows the repartition of servers between (m-d) and d then:


(2)

Under the assumptions of constant returns to scale for service rates and a perfect repartition of the number of vessel on servers, the port is therefore indifferent between a multi-user option and one to dedicate a server within the terminal. The situation is not indifferent for the shipowner calling at the dedicated and the multi-user terminal however, as long as their waiting time is higher in the second case.

The most important feature is therefore to discuss the assumptions of the previous case, in order to know when the port and the shipowners, the ones who use the dedicated and the ones who used the multi-user terminal, are not indifferent between the two situations.

The first assumption is that service times are going to remain the same when using one berth, two berths or m berths (μ_d=μ_m-d=μ_m). This assumption is challenged by some theoretical and recent empirical studies ([1], [2],[16], [17]). If economies of scale are continuous over the range of throughput, i.e. μ_d<μ_m-d<μ_m, the occupation rate for port with the introduction of a dedicated is going to decrease with the magnitude of the losses in port efficiency (service time μ). The assumption of constant economies of scale can be considered in a pure transhipment terminal where the economies of organisation implied by a better co-ordination between mother and feeder vessels could be more important that the diseconomies of scale coming from the reduction of the size of the terminal.

An others interesting case to discuss is that the total queuing waiting time can be reduced if it implies in the same time, a reduction in the variance of service [2]. In a general case, if the variance s is negatively exponentially distributed, var. (s) is s² and the global queuing time is (q) then:


(3)

If it is assumed that V(s) = 0, the waiting time can be reduced by half, and the interest of a dedicated terminal can become more accurate for the ship-owner using the dedicated terminal.

Table 1 shows, in the case of SeaLand dedicated terminal in Rotterdam. The homogeneity of ships calling at the terminal. This homogeneity and the reduction in the number of ships to be operated could lead to a routine in port operations that support the assumption of a reduction in the variance of service time.

Table 1: SeaLand' vessels calling in Rotterdam

It can be therefore considered that a better co-ordination between ships and port operations in the dedicated terminal is a way to reduce service variance and improve services (learning capacity).

The requirement by a shipping line of a dedicated terminal therefore does not automatically imply an increase in the queuing time. The previous example shows that, for the shipping line using the dedicated terminal the efficiency in terms of waiting time and through the reduction of service variance can compensate the losses induced by the decrease in service rate from m servers to d servers.

The example also underlines that losses (the increase in waiting time) could be mainly born by other shipowners moving from a multi-user terminal with m servers to a multi-user terminal with (m-d) servers. The problem with the choice of a dedicated terminal has therefore two main aspects.

a) The technical issues regarding losses and gains that could arise from the separation of terminal in different servers. b) The implication for all port users of dedicating a terminal. In order to internalise all the costs, the price for the shipowner that uses a dedicated terminal has not only to take into account its potential gains, but also the potential losses born by the other shipowners that suffer from the reduction in scale of the multi-user terminal.

In an extreme case, the best situation would be if the increase of revenue induced by the dedicated terminal gave port operators the opportunity to expand through new investments or to improve the organisation of the multi-user terminal. This solution can also be as well achieved through higher involvement of the shipping line in providing infrastructure, superstructure and management of the dedicated terminal.

A second assumption in the previous example is that ships are perfectly allocated to servers with a ratio of (d/m) and ((m-d)/m). From this assumption, the main interesting point is to make a distinction not between the number of ships (arrival rate), but between the characteristics of those ships and especially their value of time. We can for example consider the case where the queuing time in not infinite and a distinction exists between the insensitivity of ships to quit the queue according to their size. It can be therefore argued that the separation between "high speed servers" (dedicated) and "low speed servers" (multi-user) through the existence of dedicated terminals is a way for the port to keep, through differentiation, all the clients without expansion in capacity.

Conclusion

From the above discussion, it can be seen that the probability for a port to be indifferent between developing or not a dedicated terminal is low when its objective is to maximise occupation rate, under fixed prices. If one assumes that the port is maximising total revenue, it could be indifferent between the two previous situations, even if the occupation rate is lower, as long as the increase in price is going to compensate the decrease in volume. This solution could also be achieved through a higher involvement of the shipping line in providing infrastructure, superstructure and management of the dedicated terminal. At the same time, competition between ports presents a limit for this case as long as an increase in price could mean the departure or shipping lines from one port to another, in a highly competitive environment. The emergence of dedicated terminals can therefore be rather seen as a way for ports to secure futures traffics thanks to a higher involvement of shipping lines in port operations than to increase revenue through an increase in price.

Bibliography

[1] UNCTAD : Berth Throughput : Systematic Methods for Improving General Cargo Operations, TD/B/C.4/109, United Nation, 196 p., (1975).

[2] Janson J. O., Schneerson D. : Port Economics, The MIT Press, Cambridge, 183 p., (1982).

[3] Clarke R.L., An analysis of the International Ocean Shipping Conference System, Transportation Journal, 38, pp. 17-29, (1997).

[4] Hoffman J. : Concentration in Liner Shipping, ECLAC,(1998).

[4] Ryoo D.K., Thanopoulou H.A.: Liner alliance in the globalisation area: a strategic tool for Asian container carriers, Maritime Policy and Management, Vol. 26, N° 4, pp. 349-367, (1999).

[5] Meersman H., Moglia F., Van de Voorde E. : Mergers and Alliances in Liner Shipping: What does European port authorities have to fear?, Proceedings of IAME Conference, Halifax, (1999).

[6] Midoro R., Pitto A. : A Critical Evaluation of Strategic Alliances in Liner Shipping, Maritime Policy and Management, Vol. 27, n° 1, pp. 31-40, (2000).

[7] Slack B., Comtois C., Sletmo G. : Shipping lines as agents of changes in the port industry, Maritime Policy and Management, vol. 23, n°3, pp. 289-300, (1996).

[8] Heaver T. : Restructuring carriers to logistics suppliers: the challenge for international liner shipping, Canadian Transportation Research Forum, Victoria, Canada, (1994).

[9] Heaver T. : The opportunities and challenges for shipping liners in international logistics, Logistical Conference, London, Heathrow, UK, (1996).

[10] Evangelista P., Morvillo A. : Logistical integration and co-operative strategies in liner shipping: some empirical evidence, International Journal of Maritime Economics, Vol. II, No 1, January-March, pp. 1-16, (2000).

[11] De Langen P. : Time Centrality in Transport, International Journal of Maritime Economics, Vol. I, No 2, October-December, pp. 41-56, (2000).

[12] Gilman S. : The Size Economies and Network Efficiency of Large Containerships, International Journal of Maritime Economics, Vol. 1, n° 1, July-September, pp. 39-59, (1999).

[13] Cullinane K.P.B., Khanna M., Song D-W. : How Big is beautiful: Economies of scale and the optimal size of containership, Proceedings of the 1999 Halifax Conference: Liner Shipping: What's Next ?, International Association of Maritime Economists, September, pp.108-140, (1999).

[14] Saaty T.L., Elements of queuing Theory, Mc Graw-Hill, 423 p, (1961).

[15] Evans J. , Marlow P. B., Quantitative Methods in Maritime Economics, 2nd Edition, 282 p, (1990).

[16] De Monie G. : Mesure et évaluation du rendement et de la productivité des ports, Monographie de la CNUCED sur la gestion portuaire, Nations Unis, 64 p., (1988).

[17] Notteboom T., Coeck C., Van den Broeck J. : Measuring and Explaining the Relative Efficiency of Container Terminals by means of Bayesian Stochastic Frontier Models, International Journal of Maritime Economics, Vol. II, No. 2, April-June, pp. 83-106, (2000).