Vous êtes sur la page 1sur 25

Efficiency Analysis of Container Terminals in China:

an Application of DEA Approach

Bing-Lian Liu
Professor, Institute of Transportation Economics, Nankai University
ADD. 94 Weijin Road, Tianjin, China
Tel: 86-22-2350-5966; Fax: 86-22-2350-3746: E-mail: liubl@nankai.edu.cn

Wei-Lin Liu
Institute of Urban and Regional Economics, Nankai University
ADD. 94 Weijin Road, Tianjin, China
Tel: 86-22-2350-1291; Fax: 86-22-2350-3746: E-mail: liuwl@mail.nankai.edu.cn

Cheng-Ping Cheng,
Associate Professor, Department of Economics, Soochow University
ADD: 56 Sec. 1 Kuei-Yang St., Taipei 100, Taiwan
Tel: 886-952-453-336; Fax: 886-2-2382-2001: E-mail: chengche@scu.edu.tw

1
Efficiency Analysis of Container Terminals in China:
an Application of DEA Approach

Abstract

In the context of China economy’s booming up and the dominant place in the
world shipping market share, measuring the efficiency of container terminals is
important. The paper uses DEA models and Malmquist TFP approach to measure the
efficiency of China container terminals. Our data, which is from China’s port
yearbook (2004 & 2005), includes 47 terminals with an annual throughput over
10,000 TEUs, in 2003 and 2004.

Our empirical results show that the average technical efficiency scores are only
0.452 and 0.591 in 2003 and 2004 respectively. Most of the efficient terminals are
lying in the largest ports such as Shanghai and Shenzhen. The average Malmquist TFP
index score (1.202) shows that the productivity of China’s container terminals are
improved during the 2 years. Regarding the effect of capital ownership shipping line
structure characteristics, we find that the Sino-foreign Joint Ventures perform better
than Domestic Companies, and the terminals handling the containers of international
shipping lines are more efficient than those engaged in domestic shipping lines.

Keywords: Data Envelopment Analysis (DEA), container terminals, ports, efficiency

2
1. Introduction

In 2003, China has replaced America as being ranked first in total container
throughput of the world. More and more container ports, such as Hongkong, Shanghai,
Shenzhen, Qingdao, etc. have been appeared in the top list of the world container
ports. This phenomenon symbols a new age that the Chinese container ports is
playing an important role in the global shipping markets. Under the background of
globalization and China economy’s growing up, the Chinese ports now is acting as an
important gateway of China and promoting the connections between China and the
world. In the era of logistics and supply chain management when strict requests on
speed and safety are asked on the physical flow of the goods, the performance of the
ports as logistics nodes has in a large degree determined a country’s competence in
the global economy. Therefore, the efficiency of Chinese container ports is an
essential factor that is related to the competence of China.

The efficiency analysis of container ports or terminals has an important


implication for both the government and port/terminal operators. Based on the
efficiency measure, the government is able to optimize the collocation of the coast
resources and offers funding for improving the overall competence. The
port/terminal operators want to benchmark their own ports/terminals against
comparable ports/terminals to ensure competitiveness. Moreover, in the context of
continual large scale constructions are built in China’s transportation industry, it is
necessary to investigate the efficiency to avoid substantive waste of the limited
physical resources.

Although it is widespread recognition of the potential importance of ports, a


widely accepted performance measurement method for evaluating such centers have
yet to be developed (Bichou and Gray, 2004). A port is a meeting place in which
multipart organizations and institutions interplay at various levels and thus is complex.
Problems always arise when one tries to apply a single analytical instrument to a
range of ports and terminals. With the accelerating trend of containerization nowadays,

3
the container terminals have become the representation of a modern port. Due to the
significant difference in the operation processes of dealing diverse goods, the
standardization of containers has been reduced. Therefore, in recent years, it is a
significant progress to measure the technical efficiency and productivity among
various container terminals. In particular, the frontier methods that constructed upon
mathematical programming techniques have been developed with applications across
a wide range of sectors, including transit services. A recent work by De Borger,
Kerstens and Costa (2002) claims that frontier models have found their way into the
transport sector, and studies on the productivity and efficiency of almost all transport
modes are appearing. In this vein, Data Envelopment Analysis (DEA) is one of the
most important approaches to measuring efficiency.

Against this background, this paper primarily aims to estimate the relative
technical efficiency of major Chinese container terminals using 47 China’s major
terminals in 2003 and 2004, which is from China’s port yearbook (2004 & 2005).
Following DEA and Malmquist TFP approach to estimate the change of various
efficiency and productivity indexes, the non-parametric tests are further employed to
determine whether the capital and shipping line structure also affect terminal
efficiency.

The paper consists of five sections. Following introduction, section 2 investigates


the related literatures about the efficiency theory and ports/terminals. The primary
modeling techniques, DEA and Malmquist TFP approach, are briefly described in
Section 3. Data sources and the analysis of empirical results are provided in section 4.
The final section is conclusion and further research suggestion.

2. Literature Review -- Efficiency and container ports or terminals

2.1 The efficiency and container ports/terminals

Since the efficiency of a container terminal is an essential indicator of a port’s


competence for both the operation managers and government, how to measure the
efficiency and productivity becomes a key issue in port-related research. Because of
4
the complexity and multipart characteristics of a port’s activity, using a single index or
method to evaluate the overall efficiency and productivity is almost impossible.
Therefore, traditionally the approaches employed by the literature usually comprise of
a set of comparable indicators. The evaluation of port performance is mostly based
on multiple indicators. This can be indicated in the most reports of port authorities
or organizations (UNCTAD, 1976). At the same time, almost no standard method is
accepted as applicable to every port for the measurement of its performance
(Cullinane, 2003). Therefore, in order to make it possible to compare the efficiency
of different port, to construct an uniform single method that can be widely accepted
has become one of the most important objectives of the researchers recently.

In the era of containerization, the port production activity has become more
analogous than before, hence the ports that traditionally expertise in different
merchandise tend to appear a kind of homogeneity in their container terminals and
compete intensively with each other under the trend of port privatization. The
competition is not only exist in inter-port but also in intra-port. Under this
transformed market structure, DEA and stochastic production frontier method have
been occasionally used to analyze terminal production (Cullinane et al, 2005).
Thereby an overall evaluation method of their performance is attainable.

2.2 The evolution of efficiency research on ports/terminals

Over the past two decades, DEA has become a popular method for evaluating the
relative efficiencies of decision-making units within a relatively homogenous set. The
application on ports or terminals, however, is only incidental. Roll and Hayuth (1993)
is probably the first paper to discuss the applicability of DEA on port sector. It uses a
hypothetical example of 20 ports to generate simulated results. Martinez-Budria et al.
(1999) uses a DEA-BCC model to analyze 26 Spanish ports by defining them as ‘high
complexity’, ‘medium complexity’ and ‘low complexity’ ports. It comes to the
conclusion that ports with ‘high complexity’ appeared to be more efficient. In recent
years, the research on ports tends to focus more on container ports for their
homogeneous characteristics.
5
Tongzon (2001) argues that to restrict the scope of analysis to a limited number of
ports and a specific type of cargo is necessary for the multiplicity of ports and cargo
handled. It uses both DEA-CCR and DEA-additive models to analyze the efficiency
of 4 Australian and 12 other international container ports for the year 1996. This
analysis is a meaningful exploration, but the result is not pleasing for the relatively
less samples. After that, a few papers have focused on the container ports. Valentine
and Gray (2001) studies 31container ports among world’s top 100 container ports in
1998. Cullinane et al. (2005) analyze 25 container ports out of world’s top 30 ones
from 1992-99. From the literature, we find that most recently researches tend to
switch the DMUs to container terminals since individual container terminals within
ports are more suitable for one-to-one comparison than whole container ports (Wang
et al., 2002). For example, Wang and Cullinane (2006) analyzes 104 European
container terminals for the year 2003. Notteboom et al. (2000) studies 36 European
and 4 Asian container terminals, although its method is a stochastic frontier approach
rather than a DEA method.

To inspect the related literature about the research countries, most of them are
focused on study developed countries. Other than the above literature with European
terminals, only Park and De (2004) and Cullinane and Song (2003) analyze Korean
ports.

3. Efficiency Theory: DEA and the Malmquist Index

The approaches that are developed these years to measure the overall efficiencies
of the ports or terminals are evenly distributed between stochastic frontier approach
and DEA approach (Estache, 2001). Baños, Coto and Rodríguez (1999) uses a
distance function to show the difficulty faced by ports in adjusting their quasi fix
factors in the short run. Liu (1995) focuses on production frontier to calculate
technical efficiency and to compare the influence of public and private ownership in
Britain. Cullinane and Song (2003) empirically examines the relationship between the
efficiency and privatization.

6
The efficiency theory can be traced back to Farrell (1957), who identified two
types of efficiencies, technical efficiency and allocative efficiency. To measure the
technical efficiency, it is prerequisite to estimate the unobservable production frontier
either by the parametric or the non-parametric approaches. The parametric approach is
based on a specific function form and econometric methods. The non-parametric
approach is constructed directly from the sample data using linear or non-linear
programming method.

3.1 CCR and BCC models

DEA, a non-parametric method, was developed by Charnes, Cooper, and Rhodes


(1978) and further expanded by Banker, Charnes, and Cooper (1984). It provides a
valuable decision tool for the evaluation of the performance of DMUs and has been
applied in various managing control and organization diagnosis. DEA takes
advantage of the linear programming technique, without relying on the predetermined
functional forms, to construct a piece-wise production frontier over the data.
Assuming that linear combinations of the observed input-output bundle are feasible,
the convex production set that connected by efficient DMUs enveloping around all the
DMUs are estimated.

Charnes, Cooper, and Rhodes’ (CCR) efficiency concept is subject to the strong
hypothesis of constant returns to scale. Subsequent studies have considered
alternative sets of assumptions, such as Banker, Charnes, and Cooper (BCC), who
based on a variable returns to scale assumption. Banker et al (1984) assumes that
there are K firms where each DMUk (k=1,……, K) produces N outputs yik (i=1,……,
N) by utilizing M inputs xjk (j=1,……, M). The input oriented BCC model then
solve the following linear programming problems for DMU0 in each year:

7
min θ0 ,
θ0 ,λ1 ,λ2 ,L ,λK ,

s.t. - y j0 + ∑ k=1 λk y jk − W j0 + = 0,
K

θ0 xi0 - ∑ k=1 λk xik − Wi0 − = 0,


K
(1)


K
k=1 k
λ =1
λ1 , λ2 ,L , λK ≥ 0.

Where θ 0 is a scalar and λk is constants. Wi0 − and W j0 + are the input slacks and


K
the output slacks, respectively. The equation λ =1 is an important convexity
k=1 k

constraint without which the model will equal to the original CCR model. This
constraint ensures a piecewise linear frontier could be constructed.

The above model will be solved K times, once of each DMU to obtain their

respective θ value, which will be between 0 and 1. The optimal value of θ 0 is the

measure of technical efficiency of DMU0 under the variable return to scale. It is called
pure technical efficiency (PTE). If the convexity constraint is not added in, the

optimal value of θ 0 is the measure of technical efficiency of DMU0 under the

constant return to scale. It is called overall technical efficiency (OTE). The ratio of
OTE to PTE is the measure of scale efficiency (SE). The value of unity indicates a
point on the frontier and hence the it is a technically efficient firm (Farrell, 1957).
Whereas the value is less than 1, the DMU is not efficient. The value of θ means

the upper limit possible proportion that the inputs vector xi can be contracted. λk

with a positive value for DMU0 indicates the benchmarks that the firms can learn
from. Hence now we will be able to rank the DMUs according to their output and

input data given by θ 0 . We also can decompose technical efficiency into pure

technical and scale perspectives to future identify the cause of inefficiency.

3.2 The Malmquist Index

We also employ the Malmquist total factor productivity (TFP) index to measure
the impact of productivity change on the panel data. The concept of the Malmquist

8
productivity index was first introduced by Malmquist (1953), and Caves, Christensen,
and Diewert (1982) applied it in a non-parametric framework. Fare, Grosskopf, and
Lovell (1994) developed it into the output-based Malmquist productivity change
index. The Malmquist index estimates the total factor productivity (TFP) change of
a DMU between two different time periods by calculating the ratio of the distances
under a specific technology.

According to Fare et al. (1992), the Malmquist TFP index can be decomposed
into catching-up in efficiency and the shift in frontier technology. Here we define

the Malmquist productivity index by the efficiency scores of γ 0t ( X 0t , Y0t ),

γ 0t +1 ( X 0t +1 , Y0t +1 ), γ 0t +1 ( X 0t , Y0t ) , and γ 0t ( X 0t +1 , Y0t +1 ) as shown by:

1/ 2
γ t ( X t ,Y t ) ⎡ γ t + 1 ( X t + 1, Y t + 1 ) γ t + 1 ( X t + 1, Y t + 1 ) ⎤ (2)
M * (Y 1, X 1, Y 0, X 0) = 0 0 0 × ⎢ 0 0 0 × 0 0 0 ⎥
γ t + 1( X t + 1, Y t + 1 ) ⎢⎢ γ t ( X t ,Y t ) γ t ( X t ,Y t ) ⎥
0 0 0 ⎣ 0 0 0 0 0 0 ⎦⎥

The index M* is in fact the geometric mean of two input-based Malmquist TFP
indices. If M* is greater than one, then this represents positive TFP growth from
period t0 to period t1. The first term on the right-hand side of (2) measures the
technical efficiency change (EC) which represents the catching-up in efficiency. The
second term on the left-hand side of (2) measures the technical change (TC) which
represents the shift in frontier technology between period t0 and t1. If EC is larger than
one, then the efficiency improvement has occurred within the two periods. However,
if EC is smaller than one, then it indicates that terminal efficiency has become worse
than before. Similarly, if TC is larger than one, then the production technology is
progressive within two periods. The technology level is depressive if the value of TC
is smaller than one. The technical change can further be decomposed into scale
efficiency change (SEC) and pure technical efficiency change (PTEC) which is
similar with the DEA-BCC model.

9
4. Empirical Analysis

4.1 Data Sources

The rapid development of scale and dramatic structure change in Chinese


shipping industry provide a unique experiment lab for inspecting the theories and
assumptions underpinning container port production. This paper aims to evaluate
the relative efficiencies of all Chinese container terminals. Therefore, a sample
comprising all of China’s container terminals with throughput of over 10,000 TEUs
was collected. After a strict screening procedure on their reliability and integrity of
the information of each port, the sample have concluded 47 container terminals of
China for the two-year period from 2003 to 2004. The sample approximately accounts
for 83% of the total container throughput of China in each year. In sum, we collect a
total of 94 observations. The summary statistics for this sampling frame are reported
in Table 1.

Table 1 Descriptive Statistics of Inputs and Outputs Variables in the Analysis

Output Input

Quayside Gantry
Throughput Quay Length Rubber-tyred Gantry Crane
Crane
(TEU) (m) (number)
(number)
Mean 963439.4 915.12 7.14 18.93
Standard Deviation 1271140 683.70 6.62 23.31
Range 6243427 3520 38 131
Minimum 16088 230 0 0
Maximum 6259515 3750 38 131
Sum 90563303 86021.2 671 1779
Count 94 94 94 94

Our secondary data is mainly collected from China’s port yearbook (2004 &
2005). Since this published information is collected directly from the terminal
operating organizations, surveyed by an quasi-official association annually, it is
regarded as the most reliable and comprehensive data source available.

10
4.2 Input and Output Variables

The input and output variables for measuring the efficiency of container ports or
terminals industry tends to exhibit a kind of diversity in the literature, due to the lack
of uniform performance evaluation criteria. Chang (1978) suggests that the inputs of
a port should include the real monetary value of net assets in the port, the number of
laborers per year, and the average number of employees per month each year. Dowd
and Leschine (1990) argues that the productivity of a container terminal depends on
the efficient use of labor, land and equipment. In recent years, however, many
studies on container terminals tend to present a kind of uniformity. As Wang (2004),
the input and output variables should reflect the objective and process of container
terminal production as accurately as possible.

While turning to the process of container terminal production, the terminal’s


activity is to utilize the labor and equipment to accomplish the container loading or
discharging missions. In this process, the quayside gantry crane(QGC), relying on
the transfer of containers between shore and ship, is the most important equipment in
deciding the efficiency of a port (Tongzon, 1995). Before the containers are loaded
on board or after the they are discharged, to effectively handle the containers and
further serve for hinterland demand, rubber-tyred gantry crane(RTG) is another
essential equipment. During this process the ‘land’ factor can be approximate as the
total quay length(QL) of the terminals. Other input factors, such as berth working
hours, geographical position, berth waiting time and other equipments, are not
included from the consideration of both data availability and avoidance of the
problem of multicollinearity.

As to the labor input, the variable is not directly incorporated for the following
considerations, a) by analyzing the cost components, most of the cost are attribute to
the capital asserts, like fixed establishment and information technology, the salary of
the workers is very fractional. b) in the modern practice, a lot of operation during the
handling of the containers are outsourced to other companies like third party logistics
companies. Therefore the statistics data are usually not correct. c) in the era of
11
containerization, many operations of the works are standardized, the efficiency
differences caused by labor is not very significant. As is pointed out by Notteboom et
al. (2000) that a fairly stable and close relationship exists between the number of
gantry cranes and the number of dock workers in a container terminal, the labor input
could be derived by a function of the facilities of the terminal.

As to the output side, we will select the throughput as the output index in
accordance with the conventional treatment, since throughput is the most important
and widely accepted indicator for comparing the ports and terminals and also the
container is basic handling unit in the operation. Another consideration is that
container throughput is the most appropriate and analytically tractable indicator of the
effectiveness of the production of a port (Cullinane, Song and Wang, 2005). The
four inputs and the output variables are highly correlated with each other by the
Pearson Correlation Coefficients Test,, as is shown in Table 2, indicating their
complementary nature in the production process.

Table 2 Pearson Correlation Matrix of the Input and Output Variables N=94

TEU QGC RTG QL


TEU 1 0.9373** 0.9601** 0.8466**
QGC 0.9373** 1** 0.9686** 0.8250**
RTG 0.9601** 0.9686** 1** 0.8126**
QL 0.8466** 0.8250** 0.8126** 1
** Indicates significance at 1% level.

4.3 Empirical Results

As discussed previously, the DEA empirical analysis uses one output measures:
TEUs handled (the number of twenty foot container equivalent units handled) and

three input measures: QGC (quayside gantry crane), RTG (rubber-tyred gantry crane),
and QL (total quay length of the terminals). The model is base on input oriented
model in that the idle capacity loss by the expensive equipment is the main examining
purpose of this paper. The software DEAP2.1 is employed to solve the DEA models.

12
4.3.1 The DEA efficiency analysis

In order to explore the overall technical efficiency (OTE), pure technical


efficiency (PTE), and scale efficiency (SE), both the CCR and BCC models are used
to evaluate 47 container terminals in each year respectively. The three efficiency
scores calculated are given in Table 3.

13
Table 3 Terminal Efficiency under the CCR and BCC Models
2003 2004
Port Terminal
OTE PTE SE OTE PTE SE
Jinzhou Jinzhou New Age Container Terminal Co., Ltd. 0.217 0.956 0.227 0.331 0.956 0.346
Dalian Port Container Terminal Co., Ltd. 0.455 0.458 0.993 0.705 0.727 0.970
Dalian
Dalian Dagang China Shipping Container Terminal Co., Ltd. 0.394 0.591 0.667 0.502 0.854 0.588
Yingkou Yingkou Container Terminal Co., Ltd. 0.27 0.476 0.567 0.423 0.471 0.898
Qinhuangdao Qinghuangdao LG Container Terminal Co.,Ltd 0.143 0.969 0.148 0.151 0.428 0.353
Sea-Land Orient (Tianjin) Container Terminals 0.620 0.626 0.990 0.700 0.743 0.942
Tianjin Port Container Terminal Co.,Ltd 0.516 0.519 0.994 0.639 0.676 0.945
Tianjin
Tianjin Port No.1 Harbour Operating Company 0.480 0.696 0.690 0.606 0.607 0.998
Tianjin Harbour Second Stevedoing co.,Ltd 0.248 0.651 0.381 0.401 0.627 0.640
Weihai Weihai Port Container Terminal Co.,Ltd 0.130 0.541 0.240 0.187 0.541 0.346
DP WORLD YANTAI CO.,Ltd. 0.177 0.457 0.387 0.217 0.457 0.475
Yantai
Yantai Port Container Terminal 0.778 0.965 0.806 0.164 0.484 0.339
Rizhao Rizhao Port Container Terminal 0.573 0.967 0.593 0.347 0.478 0.726
Qingdao Qingdao Qianwan Terminal 0.604 0.757 0.798 0.896 0.962 0.931
Lianyungang Lianyungang New Oriental Container Terminal Co.,Ltd 0.355 0.615 0.577 0.476 0.526 0.905
Wuhan Wuhan Port Container Terminal 0.339 1.000 0.339 0.447 1.000 0.447
Zhenjiang Zhenjiang Port Group Co., Ltd. Container Branch 0.264 1.000 0.264 0.308 1.000 0.308
Nanjing Nanjing International Container Terminal Services 0.520 0.607 0.857 0.575 0.592 0.971
Nantong Nantong Port Group Co., Ltd. Container Branch 0.396 0.748 0.529 0.524 0.541 0.969
Zhangjiagang Zhangjiagang Win Hanverky Container Terminal Co., Ltd. 0.372 0.572 0.650 0.559 0.624 0.896
Yangzhou YangZhouYuan-Yang International Ports Co.,Ltd. 0.344 0.408 0.843 0.324 0.390 0.831
Shanghai Pudong International Container Terminals Co.,Ltd. 1.000 1.000 1.000 1.000 1.000 1.000
Shanghai Port Container Co., Ltd. Waigaoqiao Terminal 0.755 0.959 0.787 1.000 1.000 1.000
Shanghai East Container Terminals Co.,Ltd. 0.266 0.312 0.853 0.88 0.886 0.993
Shanghai
Shanghai Container Terminals Co.,Ltd. 0.714 0.914 0.781 0.884 0.917 0.964
Shanghai Longwu Harbour Company 0.437 0.577 0.757 0.976 1.000 0.976
Shanghai China shipping Container Terminals Co.,Ltd. 0.674 1.000 0.674 1.000 1.000 1.000
Ningbo Port Beilun Container Terminal 0.535 0.579 0.924 0.793 0.820 0.967
Ningbo Ningbo Port Beilun 2nd Container Terminal 0.568 0.572 0.993 0.730 0.740 0.986
Zhenhai Harbour Limited Company 0.234 0.827 0.283 0.320 0.832 0.385
Fuzhou Fuzhou Qingzhou Container Terminals Co.,Ltd. 0.551 0.682 0.808 0.693 0.721 0.961
Quanzhou Weitou harbour CO.,LTD 0.843 1.000 0.843 1.000 1.000 1.000
Quanzhou
Quanzhou Harbour Container CO.,LTD 0.250 0.398 0.628 0.399 0.435 0.917
Xiamen Haitian Container Terminals, Ltd. 0.504 0.509 0.990 0.526 0.589 0.893
Xiamen International Container Terminals Co.,Ltd. 0.364 0.507 0.718 0.493 0.577 0.854
Xiamen
Xiamen New World Xiangyu Terminals Co.,Ltd. 0.423 0.466 0.908 0.439 0.490 0.896
Xiamen Port Development Co., Ltd Dongdu Branch 0.378 0.925 0.409 0.438 0.648 0.676
Guangzhou Container Terminal Co., Ltd. 0.546 0.550 0.993 0.677 0.741 0.914
Guangzhou
Sinotrans Guangdong-Dongjiang Warehouse&Terminal Co. 0.23 0.494 0.466 0.34 0.486 0.700
Zhongshan Zhongshan Port&Shipping Enterprise Group Ltd. 0.481 0.497 0.968 0.678 0.840 0.807
Yantian International Container Terminals 0.739 1.000 0.739 0.723 1.000 0.723
Shekou Container Terminals Co., Ltd. 0.428 0.437 0.979 0.71 0.724 0.981
Shenzhen
Chiwan Wharf Holdings Ltd. 0.616 0.617 0.998 0.895 0.957 0.935
China Merchants Port Services (Shenzhen) Co. Ltd. 1.000 1.000 1.000 1.000 1.000 1.000
Fangcheng Fangcheng Port Terminals Co., Ltd. 0.196 0.95 0.206 0.312 0.952 0.328
Zhanjiang Zhanjiang Port China Shipping Container Terminal Co., Ltd. 0.257 1.000 0.257 0.539 1.000 0.539
Haikou Haikou Container Terminal 0.071 0.781 0.091 0.861 0.901 0.956
Average 0.452 0.705 0.672 0.591 0.743 0.791

In 2003, the average OTE score is 0.452, suggesting that a considerable


proportion of the inputs are wasted in China terminal industry. While decompose it
into the PTE and SE, their average scores are 0.705 and 0.672 respectively. They
indicate both the utilization of their existing resources and expansion on production
scale terribly need to be improved. In 2004, the average OTE score accrued to 0.591,

14
indicating that the inefficiently use of inputs is reduced. And both the PTE and SE
scores have increased to 0.743 and 0.791 respectively. This is due to the improving
management practice of the whole industry as well as to the rapid increasing demand
in the market.

Table, in 2003, 2 and 8 out of the 47 samples respectively in OTE score and PTE
score are identified as efficient. Shanghai-Pudong and China-Merchants (Shenzhen)
are identified as the most efficient terminals both in OTE and PTE perspectives. It is
not surprising since each of them is located in the largest and second largest container
ports of China. The proximity to the largest shipping market and success in
collaboration with foreign investment are the reasons behind their relative high
efficiency scores. Wuhan, Zhenjiang, Shanghai-China-Shipping, Quanzhou-Weitou,
Yantian international (Shenzhen) and Zhanjiang-China-Shipping are the six other
terminals that identified as the most efficient terminals in the PTE perspective.
Except the Yantian international (Shenzhen), which is the largest terminal in China,
most of them are quite small terminals. Their OTE scores are rather low, because of
their inefficiency in scale. This result indicates that these terminals are very busy in
the handling operation, and the devotion of sufficient resources for future expansion is
needed.

In 2004, besides the two efficient terminals in 2003, four more terminals are
ranked as the most efficient terminals in OTE and PTE perspectives. Three of them
are located in Shanghai. This advancement, on the one hand, suggests the relative
higher level of management practice in this metropolitan, and, on the other hand, is
related to government’s plan to set Shanghai as the international shipping center of
China. The other one is Quanzhou-Weitou, who located in the western shore of
Taiwan Strait. Its improvement is mainly attributed to its increase in scale
efficiency.

The terminals that are efficient in PTE scores are 11 in total, besides the above 6
efficient OTE scores the others exhibit a similar result with 2003. But it should be
noted that these terminals have all witnessed increase in their scale efficiency. This
15
improvement can be attributed to their scale expansion, which further confirmed the
analysis in 2003.

With the information about the returns to scale properties of the individual
terminal production yielded by DEA-BCC model, in 2003, 41 out of the 47 samples
exhibits increasing returns to scale, only 2 exhibits constant returns to scale. Indee,
only the four largest terminals (which lies in Shanghai, Shenzhen and Qingdao)
showed decreasing returns to scale. In 2004, however, the samples exhibiting
increasing returns to scale decrease to 26, and the samples of constant and decreasing
returns to scale mount up to 7 and 14 respectively. From this result we can infer that
with the large scale construction of the port infrastructure that happened in China, the
whole industry is evolving from an increasing returns to scale stage to a constant
return to scale stage. Although a positive relationship between the efficiency scores
and throughput still exists in 2004 (see Figure 1), this is an evidence that the effect of
scale economy that can be exploited will fade gradually in the forthcoming years.

4.3.2 The Malmquist Productivity Index (MPI)

Table 4 shows the average Malmquist Productivity Index is 1.202 which indicates
that the total factor productivity of China’s container terminals during these two years
is progressing. Among the sub-category of the index, the technical efficiency change
index (EC) is 1.34 and technical change index (TC) is 0.891. The results show a
markedly increase in the overall technical efficiencies during the two years. More
than half of the terminals’ efficiency improvements are over 1.40, and on average 1.34
is witnessed among the whole sample. This indicates a significantly catching up
among the terminals, and the difference between the efficient and inefficient terminals
are cutting down by the mechanism of mutual learning. While decomposing them
into pure technical efficiency changes (PEC) and scale efficiency changes (SEC),
although both are upswing, the improvement in scale efficiency makes the most
contribution of the technical efficiency changes.

16
The technical change (TC), however, is a little erratic since a significant inward
movement of the frontier emerged during this period, indicating the technology is
regressing while the Malmquist TFP index shown an increase in productivity. By an
inter-temporal comparison of the samples constructing the CCR frontier, we found
that although the Shanghai-Pudong and China-Merchants (Shenzhen) have all been
identified as efficient in 2003 and 2004, their performance in 2004 is not as good as in
2003. Resulting from their technical efficiency declining, the whole industry appears
a regress in technology. This phenomenon reflects that after a long-term stagnant in
the shipping market in 1990s, many port terminals are aiming at how to make more
profit during these years when the demand is rapid growing. Therefore improve
their managerial efficiency which is more directly related to increasing their profit has
a precedence over the motive of technology innovation.

Table 4 The Malmquist Index from 2003 to 2004

Port Terminal EC TC PEC SEC Malmquist

Jinzhou Jinzhou New Age Container Terminal Co., Ltd. 1.525 0.959 1.000 1.525 1.462
Dalian Port Container Terminal Co., Ltd. 1.550 0.893 1.587 0.977 1.384
Dalian
Dalian Dagang China Shipping Container Terminal Co., Ltd. 1.273 0.657 1.445 0.881 0.836
Yingkou Yingkou Container Terminal Co., Ltd. 1.565 0.866 0.991 1.580 1.355
Qinhuangdao Qinghuangdao LG Container Terminal Co.,Ltd 1.056 0.980 0.442 2.387 1.035
Sea-Land Orient (Tianjin) Container Terminals 1.129 0.899 1.188 0.950 1.015
Tianjin Port Container Terminal Co.,Ltd 1.239 0.924 1.302 0.952 1.145
Tianjin
Tianjin Port No.1 Harbour Operating Company 1.262 0.862 0.872 1.448 1.088
Tianjin Harbour Second Stevedoing co.,Ltd 1.621 0.953 0.962 1.684 1.544
Weihai Weihai Port Container Terminal Co.,Ltd 1.434 0.964 1.000 1.434 1.382
DP WORLD YANTAI CO.,Ltd. 1.227 0.910 1.000 1.227 1.116
Yantai
Yantai Port Container Terminal 0.211 0.917 0.502 0.420 0.193
Rizhao Rizhao Port Container Terminal 0.605 0.901 0.494 1.226 0.546
Qingdao Qingdao Qianwan Terminal 1.484 0.844 1.270 1.168 1.252
Lianyungang Lianyungang New Oriental Container Terminal Co.,Ltd 1.340 0.863 0.855 1.567 1.157
Zhenjiang Zhenjiang Port Group Co., Ltd. Container Branch 1.165 0.996 1.000 1.165 1.160
Nanjing Nanjing International Container Terminal Services 1.107 0.962 0.975 1.136 1.065
Nantong Nantong Port Group Co., Ltd. Container Branch 1.322 0.916 0.724 1.826 1.211
Zhangjiagang Zhangjiagang Win Hanverky Container Terminal Co., Ltd. 1.502 0.883 1.091 1.376 1.327
Yangzhou YangZhouYuan-Yang International Ports Co.,Ltd. 0.941 1.027 0.956 0.985 0.967
Shanghai Pudong International Container Terminals Co.,Ltd. 1.000 0.826 1.000 1.000 0.826
Shanghai Port Container Co., Ltd. Waigaoqiao Terminal 1.324 0.848 1.043 1.270 1.122
Shanghai East Container Terminals Co.,Ltd. 3.307 0.851 2.839 1.165 2.813
Shanghai
Shanghai Container Terminals Co.,Ltd. 1.239 0.880 1.003 1.235 1.090
Shanghai Longwu Harbour Company 2.236 0.595 1.735 1.289 1.332
Shanghai China shipping Container Terminals Co.,Ltd. 1.483 0.977 1.000 1.483 1.449
Ningbo Port Beilun Container Terminal 1.481 0.861 1.417 1.045 1.276
Ningbo Ningbo Port Beilun 2nd Container Terminal 1.286 0.920 1.293 0.994 1.183
Zhenhai Harbour Limited Company 1.368 0.887 1.006 1.360 1.213
Fuzhou Fuzhou Qingzhou Container Terminals Co.,Ltd. 1.257 0.885 1.057 1.190 1.112
Quanzhou Weitou harbour CO.,LTD 1.186 1.052 1.000 1.186 1.248
Quanzhou
Quanzhou Harbour Container CO.,LTD 1.594 0.944 1.093 1.458 1.504
Xiamen Xiamen Haitian Container Terminals, Ltd. 1.042 0.950 1.157 0.900 0.990
Xiamen International Container Terminals Co.,Ltd. 1.353 0.841 1.139 1.188 1.137
Xiamen New World Xiangyu Terminals Co.,Ltd. 1.038 0.837 1.052 0.986 0.869

17
Xiamen Port Development Co., Ltd Dongdu Branch 1.157 0.954 0.701 1.652 1.104
Guangzhou Container Terminal Co., Ltd. 1.239 0.911 1.347 0.920 1.129
Guangzhou
Sinotrans Guangdong-Dongjiang Warehouse&Terminal Co. 1.477 1.010 0.983 1.502 1.491
Zhongshan Zhongshan Port&Shipping Enterprise Group Ltd. 1.410 0.910 1.691 0.834 1.283
Yantian International Container Terminals 0.979 0.833 1.000 0.979 0.815
Shenzhen Shekou Container Terminals Co., Ltd. 1.660 0.879 1.656 1.003 1.460
Chiwan Wharf Holdings Ltd. 1.452 0.876 1.550 0.936 1.271
Fangcheng Fangcheng Port Terminals Co., Ltd. 1.595 0.960 1.002 1.591 1.531
Zhanjiang Zhanjiang Port China Shipping Container Terminal Co., Ltd. 2.094 0.641 1.000 2.094 1.343
Haikou Haikou Container Terminal 12.140 1.021 1.154 10.522 12.393
Average 1.340 0.891 1.064 1.260 1.202

4.4 Non-parametrical Test

To determine whether difference exits in various port characteristics (i.e., capital


structure, shipping line structure) for efficiency score measurement, a non-parametric
analysis (Mann–Whitney U-test) is used. These tests are executed on a PC version of
SPSS. The Mann–Whitney test is recommended for nonparametric analysis of DEA
results by Brockett and Golany (1996) and Grosskopf and Valdmanis (1987). The
final results for each characterizations are shown in Table 5 and Table 6.

Table 5 presents the results for the statistical differences on efficiency scores and
dynamic index between different capital structures of the terminal companies in 2003.
Among the 47 samples, three different kinds of companies can be identified according
to China’s port yearbook, including Companies Exclusively Funded by State,
Domestic Joint Ventures (including Share Holding Companies),and Sino-foreign Joint
Ventures. Since the operation mechanism of Companies Exclusively Funded by
State and by Domestic Joint Ventures is similar, we divide the samples into two
groups: Domestic Companies and Sino-foreign Join Ventures, each including 25 and
22 firms respectively. This division can reflect whether the reform and opening up
in the port industry have improved the operational efficiency. From Table 5, we find
that the average efficiency scores of Domestic Companies and Sino-foreign Joint
Ventures are very different. The overall technical efficiency scores of the
Sino-foreign Joint Ventures are significantly higher than that of Domestic Companies.
which is consistent with the most of previous researches. If the overall technical
efficiency is decomposed into pure technical efficiency and scale efficiency, the
Domestic Companies are better in pure technical efficiency measurement and
Sino-foreign Joint Ventures are better in scale efficiency. Therefore, the overall

18
efficiency differences are mainly attributed to their difference in scale. Since the
Sino-foreign Joint Ventures are mainly founded by multinational investors who
usually have scale advantage, they get higher scores in their scale efficiencies. In
2004, this situation of the average score is not changed in principle, but the difference
is not significant due to the catch up effect as foresaid.

From the dynamic changes of productivity during the two years, Table 5 shows
that the Domestic Terminals have improved their Malmquist TFP faster than the
Sino-foreign Joint Ventures on average. This on the one hand can be ascribe to their
effectively catching up in efficiency, especially in the scale efficiency, and, on the
other hand, ascribe to their less decline in technology. While the Sino-foreign Joint
Ventures are better in improving their pure technical efficiency, the Domestic
Terminals show the advantage of increasing their scale economic efficiency .

Table 5 Nonparametric statistical analysis about Domestic Companies and


Sino-foreign Joint Ventures

Mean Value Nonparametric statistical index

Domestic Sino-foreign Asymptotic significance


Mann-Whitney U Z
Companies Joint Ventures (2-tailed)

Score OTE 0.403 0.508 198 -1.642 0.10*

PTE 0.785 0.614 157 -2.522 0.01**

SE 0.547 0.815 124 -3.219 0.00**

Dynamic EC 1.761 1.375 242 -0.704 0.482

index TC 0.912 0.878 138.5 -2.582 0.010**

PEC 1.001 1.252 179 -2.060 0.039**

SEC 1.744 1.110 139.5 -2.889 0.004**

Malmquist 1.651 1.219 202 -1.138 0.255

* and ** denote significance at the 10% and 5% levels, respectively.

In the era of containerization, the terminals are competing with each other for the
selection to be an international shipping line by the carriers. Since the terminals that
mainly engaged in operating the international shipping line tend to have a higher
service level, a higher efficiency performance is expected. Many terminals are not
purely handling domestic or international containers, but select one as their main

19
business. We divide the whole sample into two groups according to the amount of
domestic/international containers that constitute more than 50% of the total
throughput. The 47 terminals are classified into 19 domestic and 28 international
terminals. The result of this study (see Table 6) is propitious to substantiate the
suppose above. In that, the average overall efficiency scores of international terminals
are higher than domestic counterparts, although the difference is not significant.
While decomposed into pure technical and scale efficiencies, the result is very
informative since the pure technical efficiency of domestic terminals is significantly
higher than international terminals and the scale efficiency of international terminals
is significantly better than domestic counterparts. Therefore, we can find that the
higher overall efficiency score of International Terminals is largely attributed to their
significantly higher score in scale efficiency for their relatively more substantive
investment. In 2004, relationship between the average efficiency score of the two
groups remains the same, but the difference is not significant.

While the progress of efficiencies of the two groups are examined, the
improvement of the companies that mainly manage domestics shipping lines tend to
improve faster than their international counterparts on average. It mostly attributes
to their quickly improvement in their scale efficiency. This result should be noticed
that the importance of domestic containers market is recognized by the enterprises
these years. The latest data shows that the growing rate of domestics container
throughput has run over that of domestics containers since May, 2006.

Table 6 The Nonparametric Statistical Analysis about the Shipping Line Structure of Container
Terminals

Mean Value Asymptotic significance of nonparametric analysis

Domestic International Mann-Whitney U Z Asymptotic significance (2-tailed)

Score OTE 0.414 0.478 206.5 -1.290 0.197

PTE 0.806 0.637 155 -2.412 0.016**

SE 0.543 0.760 148.5 -2.547 0.011**

Dynamic EC 1.934 1.361 224.5 -0.799 0.424


index TC 0.896 0.896 167 -1.662 0.096*

20
PEC 0.994 1.196 192 -1.520 0.129

SEC 1.893 1.170 155.5 -2.309 0.021**

Malmquist 1.833 1.216 185 -1.241 0.215

* and ** denote significance at the 10% and 5% levels, respectively.

5. Concluding Remarks

In the context of China economy’s booming up and the dominant place in the
world shipping market share, measuring the efficiency of container terminals is
important. First, in the macro level, efficiency of container terminals is an essential
factor that is related to the competence of China. The government’s planning and
funding is based on performance measures. Second, in the micro level, due to a more
drastic competition environment in a containerization era, all terminal operators need
to benchmarking themselves to find their advantage as well as shortage for improving
their competence. Based on the argument that terminal is better than port to be taken
as the unit for analysis, we uses DEA models and Malmquist TFP approach to
measure the efficiency of China container terminals. Our data, which is from China’s
port yearbook (2004 & 2005), includes 47 terminals with an annual throughput over
10,000 TEUs, in 2003 and 2004.

The empirical results uncover that there are large space for China’s container
terminals to improve their production efficiency since the average technical efficiency
scores are only 0.452 and 0.591 in 2003 and 2004 respectively. The average efficiency
scores between the two years show a significant improvement of management
practice of the enterprises. Most of the efficient terminals are lying in the largest
ports such as Shanghai and Shenzhen, indicating the large ports may have a positive
effect to the technical efficiency of their terminals for the closeness to the shipping
market. Also the result shows increasing returns to scale presented in Chinese
terminals, but the scale economy effect seems to be fading gradually. Therefore, the
efficiency will be increased if the large scale constructions of terminals speed-down in
the future.

21
The average Malmquist TFP index score (1.202) shows that the productivity of
China’s container terminals are improved during the 2 years. Among that, the
technical efficiency change index (EC) is 1.34 and technical change index (TC) is
0.891. The results show a markedly increase in the overall technical efficiencies
during the two years. However, the technical change, surprisingly indicates a regress
in technology concurrent with the improvement in Malmquist TFP index. This is
worth noticing that this result suggests that the terminal operators are more concerned
about increasing their short-term profit by expanding their scale than building their
durative competence by innovation. The shortsighted thoughts are still prevailing
among the terminal operators, especially the leading companies.

Regarding the effect of capital ownership, we find that the Sino-foreign Joint
Ventures perform better than Domestic Companies for their relatively higher scale
efficiency in 2003. The dynamic analysis shows the Domestic terminals’ catching up
are mainly focused in the scale perspective, while the Sino-foreign Joint Ventures are
founded to be more devoted in improving their pure technical efficiency.

With respect to the influence of shipping line structure characteristics on the


terminal technical efficiency, we find that the terminals handling the containers of
international shipping lines are more efficient than those engaged in domestic
shipping lines. This phenomenon is attributed to the terminals handling the
containers of international shipping lines have significantly higher scale efficiency
scores while the terminals handling the containers of domestic shipping lines show
higher pure technical efficiency. In 2004, the catch up of domestic terminals
improved their scale efficiency rapidly, resulting in a shrink of the difference of the
overall efficiency between domestic and international terminals. This also indirectly
reflects the potential of the market of domestic containers in the future.

Although the empirical results of this study have provided many useful insights
on the performance of Chinese terminals, there are still limitations needed further
consideration for the future research. First, besides the limitation of the

22
non-parametric method itself, the lack of access to stevedoring employment data

constrains the analytical power of the technical efficiency of the terminal operations.

Second, more data about the cost of various inputs are needed in order to provide
deeper insights into the operation of Chinese terminals. Third, with an extension of
the longitudinal study, the trends that exhibits in this study could be robustly inspected
by more than two years. Forth, terminals can also be classified into various clusters in
terms of size, facilities and functions (i.e., whether hub or feeder ports), and the
differences between their performances deserve to be tested.

References

Banker, R.D., Charnes, A. and Cooper, W.W. (1984) Some models for estimating
technical and scale inefficiencies in data envelopment analysis, Management Science,
Vol.30 No.9, 1078–1092.
Baños, J., Coto, P., & Rodríguez, A. (1999) Allocative efficiency and
over-capitaliation: an application. International Journal of Transport Economics, 26
(2), 181-199.
Bichou, K and Gray, R. (2004) A logistics and supply chain management approach to
port performance measurement. Maritime Policy and Management 31, 47−67.
Brockett, P.L., Golany, B., (1996) Using rank statistics for determining programmatic
efficiency differences in data envelopment analysis. Management Science 42,
466–472.
Caves, D., L. Christensen, and E. Diewert (1982) “The Economic Theory of Index
Numbers and the Measurement of Input, Output, and Productivity,” Econometrica,
Vo1 50, No. 6, 1393-414.
Chang, S. (1978)“Production Function, Productivities and Capacity Utilisation of the
Port of Mobile.”Maritime Policy and Management 5, 297–305.
Charnes, A. W. Cooper and E. Rhodes (1978) “measuring efficiency of
decisionmaking units” European Journal of Operational Research, 3, 429-444.
Cullinane K.and D.Song (2003) “A stochastic frontier model of the productive
efficiency of Korean container terminals”, Applied Economics, 35, 251-267
Cullinane, K. P. B. (2003) “The Productivity and Efficiency of Ports and Terminals:
Methods and Applications.” In Grammenos, C. T. (ed.), The Handbook of Maritime
Economics and Business, London: LLP.

23
Cullinane, K.P.B; Song, D-W. and Wang, T. (2005) A Comparison of Mathematical
Programming Approaches to Estimating Container Port Production Efficiency,
Journal of Productivity Analysis, 24, 73–92.
De Borger, B., K. Kerstens, Á. Costa (2002) Public Transit Performance: What Does
One Learn From Frontier Studies? Transport Reviews, 22 (1), 1-38.
Dowd, T. J. and T. M. Leschine (1990). “Container Terminal Productivity: A
Perspective.” Maritime Policy and Management 17(2), 107–112.
Estache, A., M. Gonzalez and L. Trujillo (2001) “Technical efficiency gains from port
reform”, Policy Research Working Papers 2637, July The World Bank
Fare, R., S. Grosskopf, and C.A.K. Lovell (1994), Production Frontiers, Cambridge:
Cambridge University Press, UK.
Fare, R., S. Grosskopf, B. Lindgren and P. Roos (1992), “Productivity Changes in
Swedish Pharmacies 1980-1989: A Non-Parametric Malmquist Approach,” The
Journal of Productivity Analysis, Vol. 3, No.1, pp. 85-101.
Farrell, M. (1957) “The measurement of productive efficiency”, Journal of the Royal
Statistical Society, Series A. Vol. 120, No. 3, 253-290
Gonzalez, M. and L. Trujillo (2005) “Reforms and infrastructure efficiency in Spain’s
container ports”, Policy Research Working Paper 3515, February. The World Bank.
Grosskopf, S., Valdmanis, A. (1987) Measuring hospital performance: a
non-parametric approach. Journal of Health Economics 6, 89–107.
Liu, Z. (1995). The comparative performance of public and private enterprises.
Journal of Transport Economics and Policy, 29 (3), 263-274.
Malmquist, S. (1953), “Index Numbers and Indifference Surfaces,” Trabajos de
Estadistica, 4, 209-242.
Martinez-Budria, E. Diaz-Armas, R. Navarro-Ibanez, M. and Ravelo-Mesa, T. (1999),
A study of the Efficiency of Spanish Port Authorities Using Data Envelopment
Analysis, International Journal of Transport Economics, XXVI (2), 237-253.
Notteboom, T, C. Coeck and J. van den Broeck. (2000). “Measuring and Explaining
the Relative Efficiency of Container Terminals by Means of Bayesian Stochastic
Frontier Models.” International Journal of Maritime Economics, 2(2), 83–106.
Park, R. and P. De (2004) “An alternative approach to efficiency measurement of
seaports”, Maritime Economics & Logistics, 6, 53-69
Roll, Y. and Hayuth, Y. (1993) Port Performance Comparison Applying Data
Envelopment Analysis (DEA), Maritime Policy and Management, 20(2), 153-161.
Tongzon, J. (1995), "The Determinants of Port Performance and Efficiency",
Transportation Research 29, 245-52.

24
Tongzon, J. (2001) Efficiency Measurement of Selected Australian and Other
International Ports Using Data Envelopment Analysis, Transportation Research Part
A: Policy and Practice, 35(2), 113-128.
Trujillo, L. Gonzalez, M. and Estache, A. (2001) Technical efficiency gains from port
reform, the potential for yardstick competition in Mexico, Policy Research Working
Paper Series 2637, The World Bank.
UNCTAD. (1976). Port performance indicators, TD/B/ C.4/131/Supp.1/Rev.1, New
York, U.S.: United Nations Conference on Trade and Development.
Valentine, V. F. and Gray, R. 2001, The Measurement of Port Efficiency Using Data
Envelopment Analysis, Proceedings of the 9th World Conference on Transport
Research, 22-27 July, Seoul, South Korea.
Wang, T. (2004). Analysis of the Container Port Industry Using Efficiency
Measurement: a Comparison of China with its International Counterparts, PhD thesis,
The Hong Kong Polytechnic University.
Wang, T., Song, D. W. and Cullinane, K. (2002) “The Applicability of Data
Envelopment Analysis to Efficiency Measurement of Container Ports,” Proceedings
of the International Association of Maritime Economists Conference, Panama, 13-15
November.

25

Vous aimerez peut-être aussi