A Study of Efficiency of Container Terminals:

A Case Study of Ports in Tanzania

Adam Ali

1,a

, I. G. N. Sumanta Buana

1,b

, I Ketut Suastika

2,c

Department of Marine Transportation Engineering, Institut Teknologi Sepuluh Nopember, Indonesia

Department of Naval Architecture, Institut Teknologi Sepuluh Nopember, Indonesia

Keywords: Container Terminals, Stochastic Frontier Analysis, Technical Efficiency.

Abstract: The objective of this study is to evaluate efficiency of container terminals in Tanzania. Relative efficiencies

between terminals are imperative to identify the potential areas of improvement for the inefficient terminals.

Tanzania is often excluded from literature about port operation and performance since many studies focused

on Asia and Europe. In order to enhance understanding about Tanzanian port efficiency, the present study is

highly demanded. However, traditional studies on container terminal efficiency tend to focus on partial

productivity measures such as TEUs per crane. These instruments do not assess the overall efficiency of

terminal operations, as they are focusing only at specific aspects of the terminal operation process. This study

uses a measurement of container terminal efficiency based on Stochastic Frontier Analysis (SFA). It is found

that the lowest score is 0.430, and the highest score is 0.997 of technical efficiency among container terminals.

On average, a typical container terminal in the sample during the study periods has an efficiency level around

0.821, meaning that the terminal operating at 82.1%, of efficiency, which is below the maximum potential

output on the frontiers. However, there is a possibility for terminals to increase efficiency by 17.9%. The most

efficient terminal ever found is at Zanzibar, and the least is at Mtwara. The urgency of automation to reduce

inefficiency level is required to fulfil the timely submission, timely delivery, and higher quality services.

1 INTRODUCTION

Seaport is a potential link of international supply

chains between sea and land transportation and

therefore enhances international trade. Following the

expansion of sea transportation technology, 80% of

world total imports and exports were conducted by

way of maritime transportation (UNCTAD, 2017)

and remains the most common mode of international

freight transport (AfDB, 2010). It is the principal

foundation to smoothing world trade, offering the

most economical and reliable way to move goods

over long distances. The growth of world trade was

due to the world container ports improvement with

enough infrastructures and handling equipment

(UNCTAD, 2018). However, many ports experience

a shortage of facilities and investment, long delays,

and dwelling time, causing congestion, affecting

import prices, and export competitiveness (Carine,

2015).

The amount of delay and dwelling time in

Tanzanian ports became significant challenges that

affected the production level due to inefficient

services provided. The efficiency of the container

terminal is an influential factor toward

competitiveness and became indicator of a country’s

development. Therefore, the Seaports authorities are

under the pressure of improving efficiency by

ensuring that the services level offered on the

container ports is on a competitive basis.

There is no doubt that technological changes

significantly influence the efficiency of container

terminals and competitiveness. The inefficiency of a

container terminal would be evidenced by several

performance indicators including physical design,

equipment, and container stacking capacity, quality

and connectivity of landsides transport connections,

links to the main shipping lines routes, vessel size,

quality of port or terminal infrastructure as well as

container handling, government process, and custom

charge. These factors are accountable in linear

relationships with economic scale since they build a

positive reputation for the customers and, indeed, lead

more attractive. Therefore, in the less matured

container port including ports of Tanzania, the study

of comparing one terminal with another in terms of

Ali, A., Buana, I. and Suastika, I.

A Study of Efﬁciency of Container Terminals: A Case Study of Ports in Tanzania.

DOI: 10.5220/0010853800003261

In Proceedings of the 4th International Conference on Marine Technology (senta 2019) - Transforming Maritime Technology for Fair and Sustainable Development in the Era of Industrial

Revolution 4.0, pages 53-62

ISBN: 978-989-758-557-9; ISSN: 2795-4579

their relative efficiency is vital for economic

reformation.

The economic value of maritime transportation

are facilitated by high-level efficiency that guarantees

timely submission, timely delivery, and high-quality

services, which are less bureaucratic. Headed to the

improvement of efficiency and productivity in the

container port or terminal, comprehensive maritime

management information systems support is needed.

These systems are automatic identification systems,

vessel traffic management systems, and port

operating systems.

In the country with a less-matured port system, the

development of port infrastructure and facilities

should be paid special attention to the port to

accommodate business activities. In Tanzania, the

ports performance has deteriorated and already

exceeded maximum capacity planned since 2013.

Therefore, container ports in Tanzania will suffer

progressive declines in operational effectiveness

unless both capacity and terminal efficiency issues

should adequately be well addressed. However, little

is known about the container terminals’ efficiency of

Tanzanian ports. This study aims to bridge the gap.

In addition, there are extensive researches in the

literature that have been conducted in examining

factors that influence performance and efficiency of

container terminal, most of them focused in Europe,

Asia, and Middle East ports (Almawsheki et al.,2015;

Zheng et al., 2016; Liu, 2010; Yang et al.,2011;

Wang, 2004), few have focused on African ports

(Ago et al., 2016; Carine, 2015). There are limited

studies focused in Tanzanian ports. Therefore, there

is a need to enhance an empirical-driven

understanding of efficiency issue to Tanzanian ports’

performance in efficiency perspective. This study

plays an essential role in creating this perspective.

The significance of this study is to provide support to

managers and operators of the container terminal in

decision making to improve the operating system in

order to produce the best potential output. It

contributes knowledge to the literature in the carrier

while helping students, researchers, and practitioners

for further development. Also, it contributes to

efficiency theories by offering an empirical model

that can be used as a decision support tool for

container terminals’ efficiency in Tanzania.

The authors designed this study with the primary

objective of evaluating the efficiency of container

terminals in ports of Tanzania with a special interest

in technical efficiency. However, approaches of

technical efficiency estimation vary depend on the

target of the researcher want to achieve. The technical

efficiency herein estimated based on output-oriented

approaches via the Stochastic Frontier Analysis.

2 LITERATURE REVIEW

Operational efficiency and effectiveness of ports or

terminals are critical to success, and considered the

best way to maintain competitiveness. Inefficient

operation and physical factors (including water depth,

mooring facilities, land, equipment, access, and so

forth) can reduce port throughput. Technological

factors have a significant impact on the availability of

real-time information for stakeholders and the

streamlining of both import and export value chains

of the business (Kahyarara & Simon, 2018). Although

port size and infrastructure, private sector

participation, and quality of both cargo-handling and

logistics services are essential determinants of

efficiency, the inputs such as quay length, terminal

area, and quay cranes have significant effects on

production (Yang et al., 2011). Similarly, the input

(length of the quay, the number of berth-side cranes,

the number of births) are shown a significant

influence on port production efficiency (Ago et al.,

2016). In the study of world major container

terminals, the variables such as terminal berth length,

alongside quay depth, terminal area, and draft, have

proven to significantly affect the production

efficiency (Hlali, 2017).

On average, container port terminals in Sub-

Sahara Africa have witnessed inefficiency indeed

rather than technical efficiency (Carine, 2015). The

study conducted at North Mediterranean Sea for both

ports and terminal operation efficiency revealed that

90% of the container ports included in the study have

their technical efficiency lower than 0.80, while 95%

of the container terminals have their technical

efficiency lower than 0.80 (Liu, 2010). From an

economic scale perspective, both container ports in

Korea and China, on average, revealed similar scores

of efficiencies about 0.886 and 0.887, respectively

(Zheng et al., 2016).

Numbers of studies have displayed different

approaches to investigate the efficiency of the ports

or terminals. Table 1 evidenced that recently, the

Stochastic Frontier Analysis (SFA) has become a

popular method in evaluating the efficiency of

container terminals. In this study, we use SFA to

evaluate the efficiency of container terminals of

Tanzania due to the available data characteristics. The

goodness of SFA is that it addresses the issue of error

and less bias as compared with analytical techniques.

Stochastic frontier analysis has not yet been used to

senta 2019 - The International Conference on Marine Technology (SENTA)

build an efficiency model of container terminals in

Tanzania ports. This situation leads to authors’

decision on using SFA approaches as paramount

techniques for studying production efficiency in

container terminals of Tanzania ports.

3 METHODS

3.1 Variables

The measurement of container terminals efficiency of

Tanzanian ports used stochastic frontier models. The

output variable that was considered in this study is

berth throughput (in TEUs) from 2010-2018. Also,

the input variables selected were quayside crane, the

terminal area in meter square, and berth length in

meter. However, other exogenous variables in binary

form were also included, such as quality of cargo

handling and private sector participation.

3.2 Modelling

The stochastic frontier analysis technique is used to

build the container terminals’ efficiency model of

Tanzania. The stochastic frontier model is a

statistical-based modelling used to analyze the

Table 1: Review of port/container terminals efficiency study.

Autho

Title Techni

ue Variables

Almawshaki et al., 2015 Technical efficiency of container terminals in

the Middle Eastern Region

DEA Berth throughput, berth length,

yard area, quay crane, yard

equipment, and maximum draft

Carine, 2015 Analyzing the operational efficiency of container

Ports in Sub-Saharan Africa

DEA Throughput, terminal area,

quayside crane, berth length,

yar

equipment

Demirel, 2012 Container terminal efﬁciency and private sector

participation

Tobit Throughput, private sector, hub

port status, logistic performance

index, and deviation distance

Hlali, 2017 The efficiency of the 26 major container ports

in 2015: Comparative analysis with different

models

SFA Throughput, quay length,

alongside depth, terminal area,

storage capacit

Hlali, 2018 Efficiency analysis with different models: The

case of container ports

SFA Throughput, quay length,

alongside depth, terminal area,

stora

e ca

acit

Liu, 2010 Efficiency analysis of container ports and

terminals

SFA Berth length, quayside crane,

yard crane, yard area, crane

spacing, trade volume, terminal

size, and throughput

Lopez-Bermudez et al.,

2018

Efficiency and productivity of container

terminals in Brazilian

orts

(

2008 - 2017

)

SFA TEUs, frequency of call, gantry

crane, an

mobile crane

Liu, 1995 The comparative performance of public and

private enterprises: the case of British ports

SFA Turnover, labour, capital,

ownership, size, capital

intensit

, and location

Notteboom et al., 2000 Measuring and explaining the relative efficiency

of container terminals employing Bayesian

Stochastic Frontie

Models

BSFM Quay length, terminal surface

area, gantry crane, and

containe

traffic in teus

Suárez-Alemán et al.,

2015

When it comes to container port efficiency, are

all developing regions equal?

SFA TEUs, terminal area, berth

length, mobile crane, and

gantry crane

Wang, 2004 Analysis of the container port industry using

efficiency measurement: A comparison of China

with its international counterparts

SFA, DEA Quay length, yard area,

quayside, and yard gantry

cranes, an

straddle carriers

Yang et al., 2011 Seaport operational efficiency: An evaluation of

five Asian port using stochastic frontier

roduction function model

SFA Berth length, quayside crane,

yard crane, yard area, and

throughput

Zheng et al., 2016 A study of container terminals efficiency of

Korea and China

DEA Berth length, quayside crane,

yar

area an

erth throughput

A Study of Efﬁciency of Container Terminals: A Case Study of Ports in Tanzania

efficiency, which identifies the frontier through the

regression method with a composed error term. The

method was first proposed by Aigner et al. (1977) and

later was improved by Meeusen and Van den Broeck

(1977), which requires the specification of

distribution assumptions in order to estimate the

efficiency. The presence of stochastic elements

makes the models less vulnerable to the influence of

outliers than with deterministic frontier models. In

general, stochastic frontier model, also called

Potential Production Function, is defined as follows:

𝑌





𝑓



𝑋





𝑒













(1)

Where:

: Observed inputs and output for an

individual container terminal

: Non-negative random variable

associated to technical inefficiency

: White noise due to random shock

The composed error terms (

and

) are

distributed independently of each other. In the

literature, the

error (

) is always normally

distributed, and (

)

is specified by several one-sided

error distributions.

The density function

can be

evaluated under the Half Normal, Exponential,

Truncated Normal, or Gamma distributions.

In this study, the Authors adopted the truncated

normal distribution assumption of Battese and Coelli

(1992) as well as Battese and Coelli (1995) models to

analyze dataset with the Cobb-Doulas function, since

the Translog function failed to accommodate the data

set accurately. The model specified in coded form

such that the first code digit represent functional form

(1= Cobb- Douglas function), second digit represents

model type (1 = (Battese and Coelli, 1992), and 2 =

(Battese and Coelli, 1995)) and the third digit

represents number of variables specification (1= three

inputs variable with exogenous variables, and 2 =

three inputs variable with trend and/or exogenous

variables).

According to Battese and Coelli (1995), the

estimation of a stochastic production frontier function

depends on the validity of variance parameters as

follows:

𝜎



𝜎





𝜎





(2)

Then, the shared variation of inefficiency is

defined as follows:

𝜎





𝜎





𝜎





𝜎





(3)

The shared variance ratio reflects the total

variation from the frontier level of output attributed

to the technical inefficiency. It is normally used to test

the null hypothesis that the technical inefficiency is

not present in the model. If that is the case the value

of variance, 𝜎





, is close to zero, and the inefficient

term must be removed in the model, and hence the

model will be constantly be estimated using the

Ordinary Least Square (OLS) method.

Furthermore, the hypothesis test for the

parameters of the stochastic production should be

diagnosed using the generalized likelihood ratio (LR)

statistic defined as follows:

𝜆2

𝐿



𝐻







𝐿



𝐻







(4)

Where:

𝐿



𝐻





: Value of log-likelihood function

restricted to OLS



𝐻





: Value of the unrestricted function

If the value of LR-statistic is significantly

asymptotically distributed as a mixed Chi-square

random variable lead the critical area with a certain

degree of freedom, the null hypothesis should be

validly rejected and potential conclusion provided.

The production model of Cobb-Douglas function in

this study is specified as follows:

ln𝑌



𝛽



𝛽



𝑋



𝛽



𝑋



𝛽



𝑋



𝛽



𝑇



𝑉



𝑈



𝑈



𝛿



𝛿



𝑍



𝛿



𝑍



𝑊



(5)

Where:

: Berth throughput of container

terminal i

: Number of quayside crane

: Area of the terminal area

: Berth length among container

terminal

: Private participation

: Quality of cargo handling

𝛽



,𝛿



: Intercept

𝛽

…

,𝛿



,𝛿



: Slope coefficients of

independent variables

: Error term needs to be estimated

T : Trend variable

, U

: White noise and inefficiency

error term, respectively

senta 2019 - The International Conference on Marine Technology (SENTA)

4 RESULTS AND DISCUSSION

4.1 Output of Container Terminals

During the period of 2010-2018, the majority of

container handling occurred in Dar es Salaam

terminal out of the total container handled, followed

by Zanzibar terminal (Figure 1). Besides, the Tanga

and Mtwara have shown relative lower operating

container trade.

Figure 1: Berth throughput 2010-2018 share.

4.2 Correlation of Variables

Correlation measures describes the relationship

between two variables. It measures the strength and

direction of linear relationships among variables. The

value of correlation in Table 2 obtained using

Microsoft Excel.

All variables are accepted since there are no

negative correlations among them. The dependent

and independent variables are reasonably correlated

and provide a venue toward analysis. whereas quay

crane has the lowest correlation with berth output.

This finding suggests relatively lower importance of

quay crane to the influence throughput of container

traffic. Among the three inputs themselves, berth

length, terminal area, and quay cranes are strongly

positively correlated to each other.

4.3 Maximum Likelihood Estimate

In general, all elasticity coefficients (beta) are

empirically found significant at a 5% level, showing

that all three inputs (quay crane, terminal area, and

berth length) have a significant effect on berth

throughput among container terminals. This result is

consistent with those observed by (Zheng et al., 2016;

Hlali, 2018; Yang et al., 2011). However, the berth

length and quay crane are not relevant since their

coefficients have a negative sign, the results are not

differently found in the study of (Lopez- Bermudez et

al., 2018 and Hlali, 2017). It is not surprising due to

sample composition in which difference of quay

crane and the length of the berth are too large among

terminals.

For both inefficient models (1.2.2 and 1.2.1), the

intercept and parameter of the exogenous variable

(private participation and quality of cargo handling)

have experienced negative signs except for the private

sector involvement in model 1.2.2. The negative sign

is indicating that private participation and quality of

cargo handling reduces inefficiency to the terminals

but not statistically significant. The result suggests that

both variables are not relevant in improving operating

efficiency among container terminals in Tanzania.

For private sector participation, it is concluded

that the container terminals can operate efficiently

without private participation. These results are proven

contrast with previous results reported by (Yang et

al., 2011; Liu, 2010; Demirel et al., 2012). These

studies evaluated efficiency significant level of

technical efficiency under private sector

participation. In the present study, Figure 3 shows

that the highest efficient container terminal is public

operating than its counterpart. The results provide

criticism for economic argument that private sector

involvement in the operation of container terminals

associated with high efficiency.

For the quality of cargo handling, the results

experienced an insignificant effect on the technical

efficiency among terminals. It means that the quality of

cargo handling is not associated with inefficiency

among terminals. However, the terminal of Zanzibar

and Dar es Salaam observed with high quality of cargo

handling, which is reflected their average efficiency

scores. There is another possibility of improving

technical efficiency among terminals if port authorities

would focus on improving the cargo handling services.

In the evaluation of container terminals, economics of

scale became a potential aspect in the running process

of any container terminal. The Authors are backing to

the production elasticity on the selected model herein,

the results displayed, thus comparing them with the

previous study.

Table 2: Correlation among terminals characters.

Y X

Y 1

0.871 1

0.967 0.951 1

0.975 0.929 0.998 1

A Study of Efﬁciency of Container Terminals: A Case Study of Ports in Tanzania

The sum of elasticity coefficients of the inputs

variables appeared to be lesser than 1, which indicates

that container terminals of Tanzanian ports shifts the

situation of constant returns to scale towards

decreasing returns to scale. The results were

supported by the study of five major container ports

that were conducted using Cobb-Douglas and

Translog function.

The summation of coefficients variable recorded

as 0.46 which is less than 1 (Yang et al., 2011).

However, the results differ from the study of

(Notteboom et al., 2000; Hlali, 2017; Suárez-Alemán

et al., 2015; Liu, 2010). The revealed behavior of

decreasing return to scale means that among the

terminals, the tendencies of using few resources of

input factors against the level of output produced

have been experienced. Therefore, the government of

Tanzania should be responsive to the port

infrastructures, investments, and policies to enhance

cargo handlings services.

In contrast with the study herein, the container

ports among 26 major ports appear to be increasing

return to scale for both model distributions (Hlali,

2017). These results suggest that 26 major ports

reached extremely usage of input factors in the

production process against the level of output

produced. The same result was observed from the

study conducted in container port of developing

countries using Cobb-Douglas function and Translog

function that tends to increased scale among the

container ports (Suárez-Alemán, et al., 2015).

However, the constant return to scale in production

process was experienced by full efficient terminals

(Almawsheki, 2015).

Management effort is required to maintain the

efficiency of handling container cargo as the results

of this study suggest that the characters of the input

among terminals are not sufficient to handle the

container cargo. Traditional inputs would surpass the

output of the production and will remain attractive to

the customers. Hence, the terminals’ authority need to

review their quality services level offered to the

customers and maintain their loyalty.

In order to decide if the model would provide

more accurate data representation in the container

terminals, several tests of the hypothesis concerning

the nature of the product function, and inefficiency

effects. The relative higher considerable value of the

log-likelihood function is satisfactory, indicating that

the model is a good fit for the dataset. This is due to

the log-likelihood is higher enough to surpass critical

value at a certain level of significance. Three null

hypotheses were assessed, and the results are

presented in Table 4.

Starting with the first null hypothesis, “There is

no technical inefficiency in the estimated model of

container terminals.” The null hypothesis was fully

rejected. That means the method used justifies the

accuracy results of the methods used.

The second hypothesis, “Technical inefficiency of

container terminals of Tanzania, is not affected by

independent variables included in the model.” This

hypothesis was also rejected, meaning that the

Table 3: Production frontier of container terminals for 2010-2018.

Variables Estimate

Parameters 1.1.2 1.2.1 1.2.2

Constant

𝛽



290.223 (0.044)* -13.793 (0.000) -40.315 (0.000)

Quay crane

𝛽



-7.385 (0.000) -4.916 (0.000) -4.641 (0.000)

Terminal area

𝛽



8.018 (0.000) 4.701 (0.000) 4.319 (0.000)

Berth length

𝛽



-10.669 (0.000) -3.204 (0.000) -2.292 (0.000)

Trend

𝛽



-0.144 (0.000) 0.012 (0.000)

Constant

𝛿



-5.411 (0.533) -2.467 (0.362)

Private participation

𝛿



-1.093 (0.797) 0.098 (0.965)

Quality of cargo handling

𝛿



-1.885 (0.634) -1.659 (0.438)

Total variance

𝜎



𝜎





𝜎





0.023 (0.000) 1.664 (0.000) 0.828 (0.000)

Gama ratio

𝛾

𝜎





𝜎





0.374 (0.000) 1.000 (0.000) 1.000 (0.000)

𝜇

0.251 (0.000)

Eta

𝜂

0.240 (0.000)

-likelihoo

16.87 15.866 17.985

Wald chi2 1667.280 2.75x10

2.99x10

* Maximum likelihood estimated parameter values obtained using STATA, at 5% level of significance with 100 iterations,

the p-value showed in the bracket. The panel data models with total observations 36 in fou

-container port terminals.

senta 2019 - The International Conference on Marine Technology (SENTA)

exogenous variables influence inefficiency among

container terminals in Tanzania.

The third hypothesis was developed to check if the

technical efficiency among container terminals in

Tanzania during the periods of study varies over time.

The postulate is full rejected since the likelihood ratio

test has been surpassing the critical area.

These hypotheses were valid to our entire models

in the study. As we have seen in Table 3, the model

specified as model 1.2.2 revealed that it is the most

correct estimation of the parameters. It is chosen as a

suitable model in this study because the value of log-

likelihood function displays higher enough than

remaining model that reflects better. Therefore, in any

piece of discussion of this study, we choose to

reference the model 1.2.2 as the best model among all

for container terminals studied and therefore

proposed to the authority of the terminals for policy

implication.

4.4 Technical Efficiency

In general, operating efficiency among container

terminals in Tanzania has shown a reasonable effort

in improving the handling of container cargo over

periods shows the pattern of terminals' efficiency

improvement across time under the study.

If technological changes effect was considered,

both Zanzibar and Dar es Salaam container terminals

are found to be gradually increasing technical

efficiency. Though, Zanzibar terminal might surpass

Dar es Salaam terminal just after 2016.

Tanga terminal, the pattern movements of operating

efficiency have showed relatively fluctuated efficient at

the beginning until 2014, in which it starts constant

decline its’ relative efficiency. Mtwara terminal, starting

with high efficiency in 2012 and start to operate

inefficiently until 2015 before starting to improve its

efficiency and surpass the Zanzibar terminal just before

2017. To conclude on the efficiency movement

observed herein, there is inconsistent with the efficiency

pattern among terminals. The ranked terminals

efficiency was also displayed.

On average, the most highly technical efficiency

terminal has been ranked, with Zanzibar terminal at

the first place, surpassing the Dar es Salaam terminal

for the substantial difference of 0.6 percent, while the

last place terminal was Mtwara terminal with worth

value 0.784 of average operating efficiency during

the nine periods. However, all terminals have

deviated far from the potential production frontier.

This result shows that during the periods under study

those terminals were not able to maximize output to

close the potential output on the frontier curve during

the production process (see Figure 2).

To compare the results with similar application in

the literature, it was found that the container port of

Shanghai, Singapore, Shenzhen, Ningbo, and Dalian

are the most efficient container ports among 26 major

ports which represent the higher number of

Containers handling (Hlali, 2017; Hlali, 2018). These

results are shown in contrast with all container

terminals in Tanzania which are almost efficient with

small number of container handling. The best

efficiency port was upholding the mean efficiency

0.876, while in the present study the best terminal was

sustained to the mean efficiency of 0.852. The results

Table 4: Hypotheses testing of the production frontier function.

Null hypothesis Log-likelihood function

Test Statistic (

)

Critical value (5%) Decision

𝐻



:𝛾0

17.985 36.380 2.706 Rejected

𝐻



:𝛿



𝛿



0

17.985 36.380 5.138 Rejected

𝐻



:𝜂0

16.870 34.150 2.706 Rejected

ote: approximate critical value at p = 5% has mixed Chi-square and obtained from Table 1 of (Kodde and Palm, 1986).

The log-likelihood function value obtained directly from the estimated maximum likelihood model (see Table 3), the

test Statistic value found from the application of Equation (5). The decision was made by comparing the difference

etween loglikelihood value and test statistics with critical area.

Figure 2: Individual technical efficiency among containe

terminals for 2010-2018.

A Study of Efﬁciency of Container Terminals: A Case Study of Ports in Tanzania

illustrate that five ports among 26 have better

management practices compared with the container

terminals of Tanzania ports.

Figure 3: Technical efficiency per terminals, 2010-2018.

The estimation of the efficiency revealed that no

single port in the sample of developing economies

had reached a full efficient input combination. The

highest-ranked port reached a technical efficiency

score of 85 percent over study periods between the

years 2000 - 2010 (Suárez-Alemán et al., 2015). The

results in supporting the results found in the present

study such that the highest-ranked terminals have

reached the efficiency of scores 85.2 percent. The

exciting results found in the Dar es Salaam port the

efficiency was relative intermediate by score 0.660,

while Tanjung Perak Port, was found lower than of

about 0.550 scores of efficiency (Suárez-Alemán et

al., 2015). It is noted that the most efficient port in

this study are San Juan - Puerto Rico, Nanjing -

China, Puerto Limón - Costa Rica, Puerto Cortés -

Honduras, Jawaharlal Nehru - India all from

developing countries while the first six ranked port

Rades of Tunisia from Africa. Note that the best

model suggested in this study is model 1.2.2, which

describes the data much more precise.

Table 5 summarises a statistical description of

technical efficiency among models. It shows that on

average a typical container terminal in the sample

during the periods has an efficiency level about 0.821,

meaning that the terminal was operating at 82.1%,

which is below the maximum potential output on the

frontier. Similarly, by holding the input factors

constant there was possibility of container terminal to

increase the efficiency level by 17.9%.

The minimum efficiency level among container

terminal is 0.430, indicating that the typical terminal

operating at 43%, which is below the maximum

potential output. There was a possibility of increasing

the efficiency by 57% if the inputs factors were held

constant. The maximum technical efficiency level

among terminals was recorded at about 0.997, which

implies that the common terminal in the sample

during the period of study operating at 99.7% close to

the maximum potential output in the frontier.

Therefore, if the terminals holding the input factors

would increase to full efficiency by 0.3%.

5 CONCLUSIONS

This study builds an empirical model under the

stochastic frontier analysis framework (1.2.2) to

study the technical efficiency of container terminals

in Tanzania ports. The model is built upon the recent

panel data covering nine years (2010-2018). The

empirical model evaluates the technical efficiency of

four container terminals. The following are

conclusions and suggestions for further studies.

5.1 Conclusions

The main findings of the study are summarized as

follows:

• Only terminal area was found to be relevant factors

of production among container terminals in

Tanzania, while berth length and quay crane did not.

• Few operating resources are still used among

terminals (decreasing return to scale), which

indicates that shortage of container handling

infrastructures faces among terminals.

• Private contribution and quality of cargo handling

are insignificant factors to technical inefficiency.

Technical efficiency among terminals in Tanzania

does not have a linear relationship with private

participation and quality of cargo handling. The

highest efficient terminal operates without private

contributions.

• As the best selected model 1.2.2, the lowest

efficiency index was 0.430, and the highest was

0.997, among terminals across the period of study.

Table 5: Descriptive of technical efficiency 2010-2018.

Model Observation Mean St

. Dev. Min Max

1.2.2 36 0.821 0.179 0.430 0.997

1.2.1 36 0.811 0.182 0.397 0.999

1.1.2 36 0.780 0.074 0.715 0.904

senta 2019 - The International Conference on Marine Technology (SENTA)

• On average, the most highly efficient terminal in

container cargo handling is Zanzibar, and the least

is Mtwara terminal.

Terminals of Zanzibar and Dar es Salaam have

emerged extremely efficient technically, even though

it is well known that they are faced with congestion.

In fact, port congestion in container terminals is

unavoidable.

5.2 Suggestions

The following suggestions are provided for future

studies:

• The effects of different ownership structures of

container terminals efficiency and productivity

should be the focus on the future research on the

container terminals in ports of Tanzania.

• In the future, the study can be extended to scale and

allocate efficiency to observe if the input resources

employed with lowest cost lead to increases the

economic profit in the container terminals.

• It is strongly suggested that in the future study, the

comparative analysis with regional container port

countries can be carried out to understand the

level of efficiency in container handling services

of each presented terminals to maintain

competitive advantage.

• In the future study, the information systems usage

and services quality levels should be a prioritized

factor to be included in the investigation of

container terminal efficiency since it currently

plays a critical role in container handling services.

• An improved estimation methodology must be

paid attention on the future study regarding

container terminals efficiency.

ACKNOWLEDGEMENT

The first author would like to thank to the Ministry of

Research, Technology, and Higher Education of the

Republic of Indonesia for being awarded the KNB-

Scholarship to help conducting this research.

REFERENCES

African Development Bank, 2010. African development

report. Ports, logistics, and trade in Africa. Oxford

University Press Inc., New York.

Ago, T. E., Hualong, Y. & Fousseni, G. M., 2016.

Container ports throughput analysis: A comparative

evaluation of China and five West African Countries

Seaports efficiencies. International Journal of

Engineering Research in Africa, Vol. 22, P. 162-173.

Aigner, D., Lovell, C. A. K., & Schmidt, P., 1977.

Formulation and estimation of stochastic frontier

production function models. Journal of Econometrics,

Vol. 6, P. 21–37.

Almawsheki, E. S. & Shah, M. Z., 2015. Technical

efficiency analysis of container terminals in the Middle

Eastern Region. Asian Journal of Shipping and

Logistics. Vol. 31, No. 4, P. 477–486.

Battese, G.E., & Coelli, T.J., 1995. A model for technical

inefficiency effects in a stochastic frontier production

function for panel data. Empirical. Economics. 20 (2),

325–332.

Battese, G.E., & Coelli, T.J., 1992. Frontier production

functions, technical efficiency and panel data: with

application to paddy farmers in INDIA. J. Prod. Anal.

3 (1–2), 153–169.

Carine, A. C. F., 2015. Analyzing the operational efficiency

of container ports in Sub-saharan Africa. Open Journal

of Social Sciences, Vol. 3, P. 10-17.

Kodde D. A. & Palm F. C., 1986. Wald criteria for jointly

testing equality and inequality restrictions.

Econometrica, Vol. 54, No. 5, 1986, pp. 1243-1248

Demirel, B. Cullinane, K., & Haralambides, H., 2012.

Container terminal efﬁciency and private sector

participation. The Blackwell Companion to Maritime

Economics, Blackwell Publishing Ltd.

Hlali, A., 2017. Efficiency of the 26 major container ports

in 2015: Comparative analysis with different models.

International Journal of Economics & Strategies

Management of Business, Vol. 10, P. 15-20.

Hlali, A., 2018. Efficiency analysis with different models:

The case of container ports. Journal of Marine Science:

Research & Development, Vol. 8, No. 2, P. 2155-9910.

Kahyarara, G., & Simon, D., 2018. Opportunity and growth

diagnostic of maritime transportation in the Eastern and

Southern Africa. University of Dar-es- Salaam, United

Republic of Tanzania, Ad hoc expertise meeting under

IAME conference.

Liu, Q., 2010. Efficiency analysis of container ports and

terminals. Thesis Ph.D. University College London.

Liu, Z., 1995. The comparative performance of public and

private enterprises: the case of British ports. Journal of

Transport Economics and Policy, 29 (3), 263–274.

Lopez-Bermudez, B., Freire-Seoane, M. J., & Gonzalez-

Laxe, F., 2018. Efficiency and productivity of container

terminals in Brazilian ports (2008–2017). Journal of

Utilities Policy, Vol. 56, P. 82–91.

Meeusen, W., & Van den Broeck, J., 1977. Efficiency

estimation from Cobb–Douglas production functions

with composed error. International Economic Review

18 (2), 435–444.

Notteboom, T., Coeck, C., & Van den Broech, J. (2000).

Measuring and explaining the relative efficiency of

container terminals by means of Bayesian stochastic

frontier models. International Journal of Maritime

Economics, 2(2), pp.83-106.

Reker, R. A., Connell, D., & Ross, D. I., 1990. The

development of a production function for a container

A Study of Efﬁciency of Container Terminals: A Case Study of Ports in Tanzania

terminal in the port of Melbourne. Australian Transport

Research Forum, PP. 209–218.

Suárez-Alemán, A., Sarriera, J. M., Serebrisky, T., &

Trujillo, L., 2015. When it comes to container port

efficiency, are all developing regions equal? Inter-

American Development Bank, working paper series

no. IDB-WP-568.

United Nations Conference on Trade and Development,

2017. Review of Maritime Transport. Palais des

Nations, 1211 Geneva 10, Switzerland

United Nations Conference on Trade and Development,

2018. Review of Maritime Transport. Palais des

Nations, 1211 Geneva 10, Switzerland.

Yang, H., Kennedy, O. R., Lin, K., & Ruth, B., 2011. Sea-

port operational efficiency: An evaluation of five Asian

ports using stochastic frontier production function

model. Journal of service science and management,

Vol. 4, No. 3, P. 391–399.

Zheng, X. B., & Park, N. K., 2016. A study on the

efficiency of container terminals in Korea and China.

Asian Journal of Shipping and Logistics, Vol. 32, No.

4, P. 213–220.

senta 2019 - The International Conference on Marine Technology (SENTA)