A Competitive Analysis of Chinese Container Ports Using the Analytic Hierarchy Process

Abstract

Over 20% of the world's container traffic occurs from Asian ports. China's entry into the market has significantly stimulated this process. Since China adopted its liberalised economic policy in the 1970s, its economy has grown at an average rate of 10% or more per annum. In particular, the efforts and investments that have been poured into its container ports are conspicuous, since approximately 90% of the country's international trade (in volume terms) is handled through maritime transport. Chinese ports (especially container ports), however, have a number of problems, such as bureaucratic administration, insufficient facilities, the lack of service and commercial orientation and inefficient operations. This paper aims to identify the competitiveness of container ports in China including Hong Kong from the outsiders’ perspective, using the framework of the Analytic Hierarchy Process, and to provide managerial and strategic implications. As expected, the findings reveal that, in terms of competitiveness, Hong Kong, Shanghai and Yantian rank first, second and third, respectively.

Appendices

APPENDIX A: TECHNICAL MATTERS OF AHP METHOD

FIRST STAGE

As a first step for the analysis, a set of problems under consideration is to be analysed and divided into a hierarchical structure. The top level of the hierarchy is an overall goal that the problems pursue. Here, the detailed elements of each level exclusive to the overall goal are 7±2 – that is, maximum allowable weights. As too much information may lead to a meaningless choice or outcome, a certain number of controllable information (in this case, ‘attributes’) is desirable (Miller, 1956; Wilkinson, 1965). Alternatives are listed in the bottom level. In principle, this is in line with the stability of the principal eigenvalue to small perturbations, when n is small and has a central role in measuring consistency.

SECOND STAGE

Pair-comparisons are to be made for the detailed elements of each level (Saaty, 1980, 1984). If n is the number of comparative elements, a decision-maker will make pair-comparisons as many times as n(n − 1)/2. The values used as a measure for the pair-comparisons are , 1/9, 1/8, ..., ½ 1, 2, …, 9 (Saaty et al, 1977; Saaty, 1980). The weights of elements at each level are computed from the pair-comparisons obtained for each level. At this time, since answers of decision-makers are not expected to be perfectly consistent, a consistency index is used to measure the degree of consistency.

To obtain weights of the criteria in the AHP method, the following is adopted: the number of criteria n is A ₁,…,A _n . If the original weights for them are W ₁,…,W _n , the comparative values of the weights of A _i band A _j (that is, a _ij ) satisfy the equation below:

Equation (A2) shows the constitution of comparison matrix A using a _ij .

When this comparison matrix A is multiplied by the vector of weights (w), vector n w is obtained as follows:

Equation (A4) is expressed in detail in (3).

Equation (A4) is manipulated for the eigenvalue and can be changed into

Here, for $w /ne 0$, n must be A's eigenvalue, when w is A's eigenvector. The eigenvalues λ _i (i=1,…,n) are all 0, with one exception. As the sum of diagonal elements is n, the only λ _i that is not 0 is λ _max. It follows as λ _i =0, λ _max=n(λ≠λ _max). Therefore, the weighted vector w for A ₁,…A _n is the normalized eigenvector (Σw _i=1) for A's principal eigenvalue, λ _max.

To solve complex problems, however, we have to obtain w′, as w is unknown. The value of w′ can be obtained by computing pair-comparison matrices, which are based on the responses of interviewees (in this case, decision-makers). Hence, the problem is altered into A′·w′=λ _max·w′, where w′ is a normalised eigenvector and λ′_max is the principal eigenvalue. In reality, the more complex a circumstance becomes, the more difficult it is to expect consistent answers from decision-makers. As such, as A′ is not consistent, λ _max always remains larger than n. This is made clear by Saaty's Theorem (Saaty, 1985) expressed in (A6):

That is, λ _max⩾n can be formed by (A6) at any time. Equality can be in place only if consistency exists. Consistency scales are shown in (A7), which is called a consistency index (CI).

When the reciprocal pair-comparison matrix A has absolute consistency, CI is 0. As CI values increase, the level of inconsistency increases. When CI is lower than 0.1, consistency is considered acceptable. If a diagonal element is 1 and symmetric elements in a matrix are in a reciprocal relationship, the average M can be obtained by a series of computations of the CI of A being randomly put 1/9, 1/8, …, ½, 1, 2, … , 9.

Equation (A8) shows how to obtain a random consistency ratio (CR) – dividing CI values computed by M, whose values are determined by Table A1:

Table ta1 Random consistency index

CR values can also be used for another index showing consistency; when the values are lower than 0.1, a solution of weighting can be considered acceptable.

THIRD STAGE

As the final step, evaluation values (in numeric terms) for elements of each level are calculated using relevant data and multiplied by computed weights from pair-comparisons. In fact, the evaluated values or numbers indicate the level of a port's competitiveness – that is, the higher the score obtained, the more competitive the port.

Appendix B: QUESTIONNAIRE SURVEY FORM

The purpose of this survey is to assess your opinion towards the relative importance of four factors related to the competitiveness of container ports in a way of pair-comparison. The four factors include cargo volume, port facility, port location and service level whose details are described below. In respect of the pair-comparison, you are requested to express which factor is more important and how important the factor is compared with its counterpart.

PART I. GENERAL INFORMATION

The four factors are extracted from the previous studies as the vital attributes to port competitiveness. The definition of each factor is given below for your reference before going through the questions. 8

Table 8 Table ta2

In making pair-comparison of the relative importance between any two factors above, the following nine scales are to be used. 9

Table 9 Table ta3

PART II. PAIR COMPARISON

Question 1

In comparing between Cargo Volume and Port Facility, which factor is more important and how important is it relative to the other factor? 10

Table 10 Table ta4

Question 2

In comparing between cargo volume and port location, which factor is more important and how important is it relative to the other factor? 11

Table 11 Table ta5

Question 3

In comparing between cargo volume and service level, which factor is the more important and how important is it relative to the other factor? 12

Table 12 Table ta6

Question 4

In comparing between port facility and port location, which factor is the more important and how important it is relative to the other factor? 13

Table 13 Table ta7

Question 5

In comparing between port facility and service level, which factor is the more important and how important is it relative to the other factor? 14

Table 14 Table ta8

Question 6

In comparing between port location and service level, which factor is more important and how important is it relative to the other factor? 15

Table 15 Table ta9