See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/304493219

Performance analysis of major ports in India: A quantitative approach

Article in International Journal of Business Performance Management · January 2016

DOI: 10.1504/IJBPM.2016.077250

CITATIONS READS

7 7,317

3 authors, including:

Anindita Mandal Jhumoor Biswas

Indian Institute of Social Welfare and Business Management Indian Institute of Social Welfare and Business Management
4 PUBLICATIONS 14 CITATIONS 38 PUBLICATIONS 369 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jhumoor Biswas on 10 August 2017.

The user has requested enhancement of the downloaded file.

Int. J. Business Performance Management, Vol. 17, No. 3, 2016 345

Performance analysis of major ports in India:

a quantitative approach

Anindita Mandal*, Soma Roychowdhury and

Jhumoor Biswas
Indian Institute of Social Welfare and Business Management,
Management House, College Square West, Kolkata – 700 073, India
Email: aninditamandal@yahoo.com
Email: srcdb@rediffmail.com
Email: jhumoorb@yahoo.com
*Corresponding author

Abstract: This paper examines the performance of 13 major ports of India in

respect of key operational performance indicators. Following rapid economic
growth India’s share in international trade is escalating. This puts increased
pressure on these ports, which handle a substantial portion of the trade to
perform with optimal efficiency. The study presents a systematic analysis of
different performance indicators for a ten-year time period (2003 to 2013)
using a variety of statistical methods and evaluates status of each port in
different categories of performance. This will enable the ports to gauge their
own effectiveness and appraise reasons for their shortcomings. In this context,
the work further develops an integrated composite performance index by
relegating comparative weightages to different indicators, to assess the relative
overall performance of different ports. The study underlines the need of such
estimates to adjudge the consistency of performance, internal and across ports
to enable planning and development of measures for enhanced performance.

Keywords: performance indicators; composite performance index; CPI;

standardised composite performance index; SCPI; port grouping; statistical
analysis; India.

Reference to this paper should be made as follows: Mandal, A.,

Roychowdhury, S. and Biswas, J. (2016) ‘Performance analysis of major ports
in India: a quantitative approach’, Int. J. Business Performance Management,
Vol. 17, No. 3, pp.345–364.

Biographical notes: Anindita Mandal completed her post graduation in

Business Administration from University of Calcutta, India. Presently, she is
doing her PhD in the Department of Business Management, University of
Calcutta. She has been working as a Systems Application Analyst in Indian
Institute of Social Welfare and Business Management, Kolkata.

Soma Roychowdhury is a Professor of Statistics at Indian Institute of Social

Welfare and Business Management, Kolkata, India and Visiting Professor of
Statistics at the University of California, Davis, USA. She has more than
25 years experience in teaching and research. She has to her credit a number of
research papers published in international journals of repute. She has authored a
number of books on probability and statistics.

Copyright © 2016 Inderscience Enterprises Ltd.

346 A. Mandal et al.

Jhumoor Biswas is an Associate Professor in Department of Environment at

Indian Institute of Social Welfare and Business Management, Kolkata, India.
She has completed her PhD from State University of New York at Albany,
USA. She has more than eight years of experience in teaching and over
15 years of experience in research on air pollution and climate change. She has
published extensively in reputed international journals and has co-authored
book chapters.

1 Introduction

Ports function as major hubs of international trade and accordingly, they are regarded as
primary catalysts of local economic development in an era of globalisation (Jung, 2011).
Approximately 90% of the international cargo is transported through ships. There are
about 4,764 ports around the world handling more than 80% of trade (ICS, 2015). Thus,
the strategic economic importance of maritime transport as a trade enabler cannot be
overemphasised. Total world port traffic has been growing at 6% to 8% per annum. The
trade competitiveness of all countries – developed and developing alike, and including
landlocked countries – depends heavily on effective access to international shipping
services and port networks. Developing countries such as China and India are major
drivers for port development due to their high economic growth rates. India is having a
large growth in international trade (over 25% compounded annual growth rate during
2003–2004 to 2008–2009). Now, 95% of India’s trade by volume and 77% by value
move through Indian ports. This trend is also true worldwide, with over 9 billion tonnes
of goods shipped internationally in 2012, and an estimated growth rate of 4.3% per year
(UNCTAD, 2013). Due to the changes in the port industry, most countries are making
great efforts to secure their ports as a hub by investing enormous funds on port facilities
and by improving efficiency in port operations and management (Lee and Kim, 2006),
which thus become an important area of study.
Data collection of key operations in a port is important as the data can be used for
enhancing port operations and provides a basis for future efficient port development. Port
roles are diverse in scope and nature. The port system serves not only as an integral
component of the transport system, but also as a major sub-system of the broader
production and logistics systems (Bichou and Gray, 2004). Ports play a vital role in the
regional economy to provide the link between suppliers and customers. From an
economic perspective, ports are increasingly related to the competitiveness of economies
(Sanchez et al., 2003). Consequent to globalisation, port performance has become
increasingly important for international trade (Bichou and Gray, 2004). The capacity as
well as the efficiency of the ports can be measured by using different performance
indicators. The indicators are filtered against specific criteria and evaluated by port
stakeholders in order to obtain a set of indicators suitable to be implemented (Puiga et al.,
2004; Bryan et al., 2007). The performance indicators are utilised to assist in
understanding of port performance trends, which can indicate actions to handle noted
situations. The performance indicators give a basis of assessment of capabilities of
management and/or operators and/or equipments, which can be improved and inculcated
in port planning measures. The measurements of port performance are applicable for port
 Performance analysis of major ports in India 347

users, policy makers, port developers, shippers and other stakeholders (De Langen and
Sharypova, 2013).
Different ports perform with various levels of efficiency. In this context, the primary
objectives of this study are to extensively review main factors that determine port
performance of the major ports in India in order to enhance efficiency of operations and
intercompare their performances to analyse their variability. Indian ports have formulated
ambitious plans for development of new ports, augmentation of existing facilities,
mechanisation of ports, purchasing of modern cargo handling equipments and
improvement in logistics to meet the challenges emerging from the anticipated growth in
trade. The capacity at 13 major ports is likely to increase to 1,459.53 million tonnes by
2020 from the level of 616.73 million tonnes in 2009 to 2010. To meet these goals, it is
imperative that ports continuously monitor and improve their functioning. Ports have
traditionally used quantitative measures to assess their performance (Marlow and
Paixao Casacaa, 2003). However, there is a deficiency of such studies conducted on
major ports of India. This study is one of the first of its kind that evaluates port
performance indicators with different statistical measures to quantify overall port
performance and efficiency. The detailed assessment of inter-comparison of performance
indicators of major ports in India is used to benchmark current best practices and identify
gaps in performances of different ports. This will indicate to the ports the degree of
improvement required in their performance in order to develop strategies to increase their
level of competence. This study develops a composite measure of overall port
performance based on amalgamation of weighted individual performance indicators to
make a holistic evaluation of relative performance of different major ports.
Port activities are usually measured by cargo output or through production function.
There are many ways of measuring port efficiency or productivity, although reducible to
three broad categories: operational indicators, factor productivity indicators, and
economic and financial indicators (Trujillo and Nombela, 1999). The present study
explained the operational indicators that refer to time measures like the ship turn-round
time (TRT), idle time (IT), pre-berthing detention (PBD) time and amount of cargo
handled by the port like average output (AO).
In order to examine the efficiency of ports a period of ten years (2003 to 2013) has
been considered for this study. It was estimated that by the end of the ninth five-year plan
of 2002, the Indian port sector required to be equipped with roughly 420 metric tons of
capacity during the ninth plan period. Nearly 70% of the additional capacity expansion
was envisaged in the major ports. In order to bridge the resource shortfall and reform
measures such as private-public partnerships were also foreseen (i-maritime, 2003).
Therefore, the analysis during this time period helps in evaluation of long-term impact of
the above policy measures.
“The operational efficiency of Indian major ports is determined primarily by
the following efficiency parameters, average output per ship-berth-day, turn-
round time, pre-berthing detention time and idle time. Every year each major
port enters into a memorandum of understanding with the Ministry of Shipping,
Government of India in which targeted efficiency parameters are agreed upon.
Performances of a port with respect to the above parameters enhance its
competitiveness.
348 A. Mandal et al.

But targets fixed by the Ministry are not based on the standard outputs of
equipments and berths and remain a mere upgrade of the previous
performances. Very low targets are set in some ports based on past
achievements which do not motivate them to achieve higher mechanization
levels and adopt better practices” (MoS, 2010).
Main objectives of Indian ports are increasing the revenue of the nation and serving its
people by fulfilling their needs, through export-import. There is a long-felt need of
developing a standardised measure by which the relative standing (as well as the absolute
standing) of different major ports can be assessed, which helps in making policy decision
for improvement of the port scenario. The inter-comparison is based on generalised
performance indicators that reflect operational efficiency. To further standardise the
measures different rates and ratios involving performance indicators have been
considered to create uniformity amongst performance of different ports for making their
performance comparable.
The data for the present study have been collected from Indian Ports Association
(IPA, 2014), Ministry of Shipping (MoS, 2014), Road Transport and Highways,
Government of India (Basic Port Statistics of India, 2011–2012). Annual figure for the
performance indicators that are effective in providing the estimates of port efficiency
have been used in this study to remove the effect of seasonality.

2 Study area

India’s long coastline is about 7,500 km of which about 5,700 km is along the mainland
and 1,800 km is around the islands. There are 13 major and about 200 minor ports. Ports
in India are classified as major or minor on the basis of ownership. The Government of
India wholly owns the 13 major ports. While the Major Port Trusts Act of 1963 governs
12 of the major ports, the 13th port, namely Ennore, is the only corporate port that is
administered by provisions of the Companies Act (1956). Major ports of India along
Indian coastline with the international codes of the 13 major ports of India together with
the abbreviations used in this article are depicted in Figure 1.
In the western coast of India, there are six major sea ports. Among them Kandla in the
State of Gujarat is a tidal port and the chief seaport serving western India. Mumbai port is
a natural deep-water harbour and is the busiest and biggest port of India. Jawaharlal
Nehru Port, also known as Nhava Sheva, is the largest container port in India and it is
the fastest growing port among the 13 major ports. Both the ports of Mumbai and
J.L. Nehru are located in the state of Maharashtra. Port Mormugao is in Goa.
New Mangalore port is located in state of Karnataka. This port serves as hinterland of
Karnataka state and to some extent the state of Kerala. Cochin port is a major port on the
Arabian Sea in Kerala. In eastern cost of India, there are seven major ports. Among them
Kolkata and Haldia are both riverine ports and others are sea ports. Kolkata port situated
in the bank of Ganges in state of West Bengal is the oldest operating port in India. The
Kolkata Port Trust (KoPT) manages two separate dock agglomerations – the Kolkata
Dock System and the Haldia Dock Complex. Paradip port is the primary port in state of
Odisha. Visakhapatnam port is the only major port of the state of Andhra Pradesh. It is
the deepest land-locked and protected port in India. Chennai port is one of the oldest
major ports in India in the state of Tamil Nadu. Ennore port (officially renamed as
Kamarajar port since February 26, 2014) is located on the Coromandel coast about
 Performance analysis of major ports in India 349

24 km north of Chennai port in Tamil Nadu, and is the first port in India, which is
a public company. Tuticorin port, now known as V.O. Chidambaram port, located in
Tamil Nadu, is the southernmost port in India.

Figure 1 Major ports of India with abbreviation

3 Methodology

Due to the ongoing commercialisation of port authorities (PAs) and the increasing
pressure of stakeholders on PAs, new port performance indicators are developed. Such
new indicators do not only satisfy the port authority’s need for insight in port
performance, but are also relevant for stakeholders with socio-economic interests in a
port (De Langen et al., 2007). We have developed a new concept of composite
performance index (CPI) in this study to indicate overall performance efficiency of ports.
In addition, we have modified and then utilised performance indicators to evaluate
internal consistency and conduct inter-port analysis.
The operational efficiency of Indian major ports is determined primarily by the
following efficiency parameters (MoS, 2010): AO per ship-berth-day, TRT, PBD time
and IT, which are defined as follows.
AO per ship-berth-day is the AO (for loading and unloading) of a ship at a berth per
day. It is measured in tonnes of cargo.
TRT indicates the average time spent (in days) by a ship since its entry into the port
till its departure. It is used to evaluate the ship efficiency in a port and is mainly
influenced by the cargo volume.
Since TRT is affected by the volume of the cargo, to get a better indicator of
efficiency, one should consider TRT with respect to the output of the port (in tonnes).
350 A. Mandal et al.

Thus, a new measure is proposed in this study, average turn-round time (ATRT), defined
as follows:
TRT
ATRT = × 1, 000,
AO
measured in days per thousand tonnes.
PBD time indicates the time (in days) for which a ship waits before getting entry into
a berth. But PBD of two ports are only comparable if their outputs per berth are similar.
Instead, it is more meaningful to consider the ratio of PBD to AO as an improved
indicator. Hence, a modified measure, detention time to output ratio (DTOR), is proposed
here, which is defined as follows:
PBD
DTOR = × 1, 000,
AO
measured in days per thousand tonnes.
IT at berth signifies the time when a vessel remains idle at berth. It is expressed as a
percentage of the total time a vessel takes at berth.
Previous research work has used operational parameters, primarily TRT, to evaluate
port performance (UNCTAD, 1976; Kasypi and Muhammad, 2006). Present paper adopts
more comprehensive methodologies since it utilises all operational indicators and
modifies them according to the approach stated above to maintain consistency among
port for inter-comparison.
The measure of performance of the above factors can help in identification of
potential bottlenecks in port logistics and assessment of performance, capacity and
service levels of ports. The first performance indicator can directly translate into revenue
earned by port. The other performance indicators convert to total time in port per ship and
the cost of time per ship that render voyage cost for a ship. Managers in various port
terminals try to attract carriers by speeding up various services, and reduce costs by
utilising resources efficiently.
In the present work, a CPI is developed, by assigning relative weightages, to the
indicators, AO, ATRT, DTOR, IT, to assess the overall performance of different ports, as
described below.
Let wi be the weight of indicator i, yij be the data-value of ith indicator for jth port,
i = 1, 2, …, k, j = 1, 2, …, c. Thus, for a particular year, we obtain the CPI, expressed as
percentage, for jth port, using the following expression:

1
⎧ k
⎪ ( ) ( )
⎛ yij − min yij × I1[i ] + max yij − yij × I 2[i ] ⎞ ⎫
⎜ ⎟⎪
∑
j j
CPI ( j ) = ⎨ wi × ⎜ ⎟⎬
k
⎪ i =1 ⎜ max yij − min yij ⎟⎪
∑i =1
wi ⎩ ⎝ j j ⎠⎭
(1)

× 100, j = 1, 2, ..., c,

where the indicator functions I1[i] and I2[i] are as follows:

⎧1, if larger value of yij indicate better performance, for the i th indicator
I1[i ] = ⎨ th
⎩0, if smaller value of yij indicate better performance, for the i indicator,
 Performance analysis of major ports in India 351

and

⎧1, if smaller value of yij indicate better performance, for the i th indicator
I 2[i ] = ⎨ th
⎩0, if larger value of yij indicate better performance, for the i indicator.
Note that, for AO, the bigger the value of yij, the better is the performance, whereas for
the rest of the indicators, the reverse is true. A port performing best in all categories will
get a score of 100%, while a port having worst performance in all categories will get a
score ‘0’, according to the expression of CPI, as given in (1).
For assessing the relative standing of different ports with respect to their overall
performance, we propose a modified score function, standardised composite performance
index (SCPI), as follows:
CPI ( j ) − mean(CPI )
j
SCPI ( j ) = , j = 1, 2, ..., c, (2)
sd (CPI )
j

for jth port, the value of which can be interpreted as follows:

SCPI = 0 indicates an average performance, SCPI > 0 indicates above average
performance, and SCPI < 0, below average performance. To be more specific, SCPI > a
(a > 0, a real constant) means the CPI to be ‘a’-times above standard deviation of all
CPIs, and SCPI < –a, to be ‘a’-times below standard deviation. Thus SCPI-value reveals
the extent of quality of performance for each port, for any given year. It indicates the
relative standing of each port, as well as their absolute position.
As CPI, a score function, will be used to generate rankings, it should possess an
important property of being invariant to monotonic transformations.
Let R(j) be the overall rank of port j, derived from CPI, and R ( j ′) be the overall rank
of port j ′. Suppose, R ( j ) ≤ R ( j ′). Smaller rank indicates better overall performance.
Note that the above indicator CPI, as given in (1), satisfies the property of a score
function that it is invariant to monotonic transformations, i.e.,
R( j ) ≤ R ( j ′)
⇔ CPI ( j ) ≥ CPU ( j ′ )
⇔ g ( CPI ( j ) ) ≥ g ( CPI ( j ′ ) ) ,

where g is a monotonically increasing function.

This property of the score function implies that the rank of the ports will remain same
if they are obtained according to another increasing function g of CPI. Here SCPI, as
defined in (2), is an increasing function of CPI. Hence, ranking through SCPI gives the
same sequence obtained through CPI. Note that the SCPI also holds the same invariance
property of score function.
Here, the effect of time from the data on time-sensitive indicators is removed to make
them comparable at any particular point of time through necessary statistical tests as
elicited below for comparing equality of means of different ports in respect of each of the
listed indicators (AO, ATRT, and DTOR) and the composite index (CPI/SCPI).
The proportion of time a port exceeded its own standard, for each of the indicators, is
also needed to measure its own consistency over the years. From the variability of the
data, an idea can be formed on how consistent a particular port has been over time.
352 A. Mandal et al.

Together with mean and standard deviation of the data values, coefficient of variation
(cv) has also been determined to assess and compare the relative dispersion of the data
values for different ports. A 95% confidence interval, based on t-statistic, for mean has
been determined for different ports to get an idea about the possible range of their
average values. As a measure of exceedance limit, here (mean – sd) is regarded as the
lower limit in case of AO, and (mean + sd) as the upper limit for the others. Note that for
any data coming from a normal population, the percentage of cases of exceedance should
be no more than 15.87%. It will be alarming if this value exceeds the given limit, for a
particular port.
Here, we have used box plots to compare the characteristics of the distributions of
each of the indicators in respect of their minimum value, lower quartile, median, upper
quartile and maximum value. It also displays the outliers in a dataset, if any.
The method of agglomerative hierarchical clustering with average linkage has been
used here for grouping ports that are similar with respect to different performance
indicators. In this analysis, the Euclidean distance has been used as a proximity measure,
where an Euclidean distance between ports j and j ′ is given by
1
⎧⎪ k 2⎫
⎪2
d jj ′ =⎨ ∑
⎩⎪ i =1
yij − yij ′ ⎬ ,
⎭⎪
j , j ′ = 1, 2, ..., c, (3)

where k is the number of performance indicators, c is the total number of ports under
study.
Since clustering technique is not capable of grading or ranking the ports according to
their performance in different categories, other statistical techniques such as F-test, t-test,
Kruskal-Wallis test are used for determining finer differences among the groups and
grading the ports accordingly.
On the basis of the results of the above mentioned tests, ports are grouped on the
basis of similar performance in each category of performance measures and the
composite index as well. Note that grouping of similar performing ports using ANOVA
based on CPI and SCPI will be the same, as they are linear function of each other.
Kolmogorov-Smirnov test, a non-parametric test, has been carried out to examine the
normality of the data so that ANOVA can be performed. Kruskal-Wallis test is performed
in case of IT, as it is expressed as percentage, for which a parametric test based on
ANOVA is not applicable.

4 Results and discussion

There are limited studies using operational performance indicator in the Indian scenario
regarding ports. The most recent such study (Deshmukh, 2011) employs descriptive
statistics for inter-comparison of port performances. The present study improves upon
this by first ensuring consistency amongst the operational parameters as discussed in the
previous section and devising a CPI that present a holistic performance of port
operations. The limitation of this study is that it can only explain major discrepancies
amongst the ports.
 Performance analysis of major ports in India 353

4.1 Composite performance index

In the present work, we develop a CPI, by assigning relative weightages to the indicators,
AO, ATRT, DTOR, IT, to assess the relative overall performance of different ports year
wise. A questionnaire on assigning relative weightages to different performance
indicators was distributed to five experts, who have long experience and expertise on port
operations (Delphi method, Dalkey and Helmer, 1963). Each of them was asked to assign
relative weightages to all possible pairs of indicators. Their responses were redistributed
among themselves, just in case they wanted to revise their responses. All experts were
anonymous to each other. Their final opinion on relative weightages were considered and
averaged. Then analytic hierarchy process (Saaty, 1980) was applied to the opinion
received to determine the priority vectors (normalised principal eigen vectors) which
finally give the weightages of different indicators. It has been seen that Ennore is doing
well consistently over the years with a high average CPI for 2003 to 2013, followed by
J.L. Nehru. The SCPI of Ennore port is uniformly greater than 1, it is followed by J.L.
Nehru and New Mangalore.
Other ports like Paradip, Visakhapatnam, Mumbai, Kandla, Cochin, Mormugao and
Chennai have mixed performances over the years. But ports like Kolkata, Haldia and
Tuticorin consistently perform below average for all years with negative values of SCPI.
The CPI and SCPI-scores for 2003 to 2013 are shown in Table 1. The table reflects an
improvement or deterioration in overall performance of the ports for the study period.
In 2012 to 2013, CPI of J.L. Nehru was the highest (97.81%), which is 1.69 times
standard deviation above mean, as determined by the SCPI-score, while respective CPIs
of KDS and HDC were as low as 15% and 20.49%, which were –1.85 and –1.61 times
standard deviation below mean, respectively. Overall performance of Mormugao was
average (mean CPI) of all ports, as indicated by its SCPI-score. HDC and Ennore showed
deterioration in 2012 to 2013 from their previous CPI, while all other ports showed an
improvement. The CPI developed in this study precisely reveals the overall performance
of the ports and is found to be a robust index of measurement of port performance on the
whole for any period of time.

4.2 Evaluation of internal regularity of ports

In this study, an internal regularity of ports in different performance indicators are
checked. A benchmark has been set across all performance indicators for every port to
evaluate the deviations from the standard (mean) as well as from a limit (mean ± sd). An
internal consistency check on all parameters will enable a port to assess the extent of
fluctuations in data and accordingly try to identify the causes in order to develop a
strategy to mitigate the problem for optimising operational performance. Here, for each
port, we determine the percentage of times the port has performed below (mean – sd) or
above (mean + sd), as the case may be, for each performance indicator. In a particular
time period, a value going beyond that limit indicates an alarming performance under the
specified category. We find that in 30% cases the AO of Mormugao has gone below
(mean – sd) during the study period. The ports of Ennore, Tuticorin, Cochin and
Haldia have performed below their own limit in 20% cases. As seen in Table 3,
the performances of the ports like Paradip, Visakhapatnam, Kolkata, Chennai,
New Mangalore, J.L. Nehru and Mumbai are below their limits in 10% cases.
 354

Table 1

CPI (%)
A. Mandal et al.

Port
2003–2004 2004–2005 2005–2006 2006–2007 2007–2008 2008–2009 2009–2010 2010–2011 2011–2012 2012–2013
Kolkata Dock System 5.73 6.94 9.08 12.73 7.50 7.69 0.43 6.38 3.67 15.00
Haldia Dock Complex 30.43 25.99 22.54 19.13 24.77 27.59 31.72 35.67 24.92 20.49
Paradip 42.02 40.50 40.41 40.44 34.78 50.40 59.28 66.72 57.31 69.86
Visakhapatnam 45.00 43.09 41.13 40.80 43.57 51.43 59.84 56.31 42.87 52.12
Ennore 100.00 100.00 99.05 99.14 97.17 97.23 95.23 88.08 100.00 93.73
Chennai 26.55 35.50 40.88 38.82 38.53 48.47 62.27 63.16 50.77 61.31
CPI and SCPI values of different ports

Tuticorin 23.41 24.09 22.72 9.67 31.31 31.22 46.46 48.56 30.67 41.06
Cochin 36.38 35.09 36.05 35.60 47.65 52.81 64.78 68.16 67.27 72.39
New Mangalore 56.25 55.26 56.71 50.32 51.81 62.54 73.77 78.20 64.85 72.62
Mormugao 47.91 49.08 54.61 52.84 36.87 30.33 26.87 17.19 41.35 58.85
J.L. Nehru 55.54 59.27 63.46 63.77 67.25 81.06 99.10 99.37 74.48 97.81
Mumbai 22.94 22.71 23.61 6.36 23.14 29.93 47.07 46.77 40.54 47.90
Kandla 28.69 27.39 30.40 25.38 39.08 50.61 69.35 73.82 54.43 53.96
 Table 1

SCPI
Port
2003–2004 2004–2005 2005–2006 2006–2007 2007–2008 2008–2009 2009–2010 2010–2011 2011–2012 2012–2013
Kolkata Dock System –1.55 –1.51 –1.46 –1.04 –1.61 –1.76 –2.17 –2.00 –2.00 –1.85
Haldia Dock Complex –0.44 –0.65 –0.86 –0.78 –0.80 –0.89 –0.96 –0.86 –1.09 –1.61
Paradip 0.09 0.01 –0.05 0.10 –0.33 0.11 0.10 0.36 0.30 0.50
Visakhapatnam 0.22 0.12 –0.02 0.11 0.08 0.16 0.12 –0.05 –0.32 –0.26
Ennore 2.71 2.70 2.58 2.50 2.61 2.17 1.49 1.19 2.14 1.52
Chennai –0.61 –0.22 –0.03 0.03 –0.15 0.03 0.22 0.22 0.02 0.13
Tuticorin –0.75 –0.74 –0.85 –1.16 –0.49 –0.73 –0.39 –0.35 –0.84 –0.73
Cochin –0.17 –0.24 –0.25 –0.10 0.28 0.22 0.31 0.41 0.73 0.60
Performance analysis of major ports in India

New Mangalore 0.73 0.67 0.68 0.50 0.47 0.65 0.66 0.81 0.63 0.61
CPI and SCPI values of different ports (continued)

Mormugao 0.35 0.39 0.59 0.60 –0.23 –0.77 –1.15 –1.58 –0.38 0.03
J.L. Nehru 0.70 0.86 0.98 1.05 1.20 1.46 1.64 1.63 1.04 1.69
Mumbai –0.77 –0.80 –0.81 –1.30 –0.88 –0.78 –0.37 –0.42 –0.42 –0.44
Kandla –0.51 –0.59 –0.50 –0.52 –0.13 0.12 0.49 0.63 0.18 –0.18
355
 356

Table 2

95% lower 95% upper Coefficient of % below % below

Port Mean sd (mean – sd)
confidence limit confidence limit variation (%) mean (mean – sd)
A. Mandal et al.

KDS 3,881.25 601.58 3,508.39 4,254.11 15.50 70% 3,279.67 10%

HDC 8,774.45 618.74 8,390.95 9,157.95 7.05 50% 8,155.71 20%
Paradip 9,671.25 620.68 9,286.55 10,055.95 6.42 50% 9,050.57 10%
Visakhapatnam 10,497.59 259.95 10,336.47 10,658.71 2.48 50% 10,237.64 10%
Ennore 36,087.10 4,102.95 33,544.07 38,630.13 11.37 40% 31,984.15 20%
Chennai 9,352.55 535.64 9,020.56 9,684.54 5.73 40% 8,816.91 10%
Tuticorin 4,617.16 343.83 4,404.05 4,830.27 7.45 40% 4,273.33 20%
Cochin 6,161.50 1,002.40 5,540.20 6,782.80 16.27 50% 5,159.10 20%
New Mangalore 13,712.30 895.90 13,157.02 14,267.58 6.53 60% 12,816.40 10%
Mormugao 15,885.02 4,068.53 13,363.32 18,406.71 25.61 50% 11,816.49 30%
J.L.Nehru 13,128.00 1,622.65 12,122.27 14,133.73 12.36 50% 11,505.35 10%
Mumbai 4,988.25 936.75 4,407.65 5,568.85 18.78 60% 4,051.50 10%
Kandla 7,910.70 602.34 7,537.37 8,284.03 7.61 60% 7,308.36 10%
Evaluation of internal consistency of ports with respect to AO*

Note: *In thousand tonnes

 Performance analysis of major ports in India 357

Different statistics presented in Table 2 together give an overall idea about the internal
regularity in AO of each port. The lower and upper limits of a 95% confidence interval of
mean have been computed. The cv indicating the relative dispersion of AO for each port
has also been presented. Visakhapatnam is found to be the most consistent in AO (with a
cv of 2.48%), while Mormugao has been most inconsistent (with a cv of 25.61%) among
all ports, during 2003-2013.
Similarly, for ATRT (table not shown), Tuticorin, Cochin and Visakhapatnam have
crossed their own limit of (mean + sd) in 30% cases. For the ports like Ennore, Paradip,
Chennai, New Mangalore, Mormugao, J.L. Nehru and Kolkata have exceeded the limit in
20% cases. The limit is exceeded in 10% cases for Haldia, Mumbai and Kandla.
In case of DTOR, New Mangalore has never exceeded its limit throughout the study
period. Then come Chennai, Cochin and J.L. Nehru with exceedance in 10% cases. The
remaining ports have crossed their limits in 20% cases.
In case of IT, Tuticorin has never crossed its limit throughout the entire study period.
Cochin, New Mangalore, Mumbai and Mormugao have crossed their limits in 10% cases.
Paradip, Visakhapatnam, Ennore, Kolkata, J.L. Nehru and Kandla have exceeded the
limit in 20% cases. Haldia and Chennai have exceeded their own (mean + sd) in 30%
cases.
Overall, assessing each port’s exceedance of its own limit in various categories of
performance, New Mangalore has the lowest number of deviations, while Haldia,
Vishakapatnam, Ennore and Mumbai have the maximum.

4.3 Box plot

Different box plots are placed side by side in a same plot, along with a box plot for the
average over all ports taken together, enabling us to compare the summary statistics for
different ports.
In box plot of AO [Figure 2(a)] of different ports, Ennore has been found to be best
performing among all major ports. The minimum value of AO for Ennore is above
28 thousand tonnes during the study period. The maximum value of AO for other ports is
less than the minimum value of Ennore.
In case of IT [Figure 2(b)], New Mangalore, Mormugao, Ennore and J.L. Nehru are
among the best performing ports with values ranging between 7% and 24%
approximately, while for Paradip, Visakhapatnam, Chennai, Cochin and Tuticorin, the
values span from 25% to 35%. The poorly performing ports are Kolkata and Haldia, with
a maximum of 40% and 37%, respectively.
As clearly visible from [Figure 2(c)], Ennore and J.L.Nehru are the best performing
ports in the category of ATRT, with a maximum of 0.14–0.15 days/thousand tonnes.
Again, Kolkata port fairs poorly in this category ranging between 0.85–3.6 days/thousand
tonnes.
358 A. Mandal et al.

Figure 2 Boxplots showing the distribution of (a) AO (b) IT (c) ATRT (d) DTOR

(a)

(b)

Notes: o suspect outlier, * extreme outlier.

 Performance analysis of major ports in India 359

Figure 2 Boxplots showing the distribution of (a) AO (b) IT (c) ATRT (d) DTOR (continued)

(c)

(d)

Notes: o suspect outlier, * extreme outlier.

360 A. Mandal et al.

Ennore is the best performing port in the category of DTOR, followed closely by
New Mangalore, J.L. Nehru and Cochin. Haldia and Kolkata do not perform well in this
category. The minimum value for Haldia is as high as 0.14 days/thousand tonnes, while
the maximum for Ennore is 0.08 days/thousand tonnes.
Amongst all outliers seen in the box plots, IT of Tuticorin in the year 2007 is
extremely low, as compared to its performance in the other years, while ATRT and
DTOR of Mormugao in the year 2010 are extremely high.
Overall, Ennore exhibits best performance followed by J.L. Nehru. The poorly
performing ports need to strive to attain the benchmarks set by the best performing ports.

4.4 Cluster analysis

Clustering technique has been applied on observed values of different performance

indicators of all major ports to divide them into disjoint groups. Any ports belonging to a
same cluster are similar in their values under different categories of performance, while
ports in different clusters are dissimilar. Similarities are measured using Euclidean
distance, as given in (3). The results are displayed in a dendrogram, as shown in Figure 3.

Figure 3 Dendrogram showing clusters of ports

 Performance analysis of major ports in India 361

From the dendrogram, we can identify three clusters as under:

Cluster 1: E
Cluster 2: JLN-NM-Mo
Cluster 3: V-P-Ch-Co-HDC-Ka-T-Mu-KDS.
As clusters produce only raw groups for further finer grouping of ports we perform some
statistical tests, which are discussed in the next subsection.

4.5 Grouping of ports

Kolmogorov-Smirnov test conducted on performance indicators reveals that the p-value
is more than 0.05 for all cases indicating normality of the populations. An F-test based on
ANOVA has been carried out to test whether the mean performance for all ports are same
in each category of performance indicators (AO, ATRT, DTOR) and for CPI. Note that
ANOVA based on CPI and SCPI will produce the same result, as they are linearly
related. In case the null hypothesis of equality of all 13 port means is rejected, we
proceed for t-tests to examine the equality of means pairwise. As the indicator, IT, is
expressed as percentage, we carry out Kruskal-Wallis test, a non-parametric test used for
testing equality of medians. If the test rejects the null hypothesis of equality of all port
medians, we compare them pairwise to detect which of the pairs are similar having
insignificant difference in their ITs. This way we can group similar ports, the differences
in central values of which are statistically insignificant.
Note that the groups may not be disjoint. Any ports belonging to the same group do
not have significant difference in their means. The ports belonging to two different
groups have significantly different means. But there are some ports that belong to more
than one consecutive group. This indicates that the difference in means of these ports are
not significant from the means of the ports belonging to any of those groups, but the other
ports belonging to different groups have significantly different means.
We introduce a plot, called chain plot (Figure 4), to display the inter-comparison of
groups of ports which are similar in relation to different indicators. It is a summary of the
results obtained from above statistical tests (F-tests based on ANOVA, t-tests,
Kruskal-Wallis test), and the plot indicates relative positions of individual ports in
different categories of performance indicators. The test results reconfirm the earlier
findings. The CPI/SCPI accurately reflects the overall status of the ports at a glance since
it corroborates earlier results and can be effectively used to analyse overall standing of a
port during any time period. Overall, Ennore is the best performer followed by J.L. Nehru
and New Mangalore.
The hierarchal clustering technique applied in this study successfully allots the
different ports in similar performing groups based on different performance indicators.
Ennore, the best performing port, with three berths that handle mostly coal; and one
terminal dedicated to petroleum, oil and lubricants (POL) and other types of liquid cargo.
This port has scope for increasing traffic growth by optimising its capacity handling
processes such as berth occupancy. The next two best performing ports J.L. Nehru and
New Mangalore have nearly equal number of berths. J.L. Nehru mostly handles container
traffic and New Mangalore handles general break-bulk cargo and liquid bulk. Both
Visakhapatnam and Mormugao handle mostly general break bulk cargo and liquid cargo.
362 A. Mandal et al.

The next similar type of port group consists of Paradip, Chennai, and Cochin with equal
number of berths. Both Chennai and Cochin handled similar type of cargo, container and
general break bulk cargo, but Paradip handled all types of cargo except container. Among
these three ports Paradip is the best, which is also reflected in the relative position of
Paradip in the chain plot. Haldia and Kandla are in a similar group followed closely by
Tuticorin and Mumbai. Haldia and Tuticorin both specialise in general break bulk cargo
and liquid bulk cargo. Kandla and Mumbai are the largest ports and handle all types of
cargo. Kolkata handled mostly general break bulk cargo, container and liquid bulk cargo.

Figure 4 Chain plots showing grouping of ports based on their performance in various categories
 Performance analysis of major ports in India 363

5 Conclusions

A comprehensive study for the time period (2003 to 2013) has been conducted for
operational efficiency indicators for the 13 major ports in India. The primary motivation
of this work is that in the current scenario of rapid economic development, Indian ports
face a challenging situation to enhance their performances and need benchmarks to
accomplish that. It is important for ports to attain benchmarks in performance for
continual improvement. Annual performance indicators have been considered in this
study to remove effect of seasonality.
Grouping according to the overall indicator based on CPI/SCPI successfully reflect
the influence of operational parameters. Major ports of India handled in excess of
capacity, as a result of which TRT is very high and productivity is low in comparison to
other ports of world. Indian ports are also not retrofitted to handle new types of cargo.
This type of assessment will enable ports to evaluate their efficiency and incorporate
methods to improve their deficiencies to at least meet internal benchmarks. This is a
necessary first step for developing policy measures to enable Indian ports to become
globally competitive and strive towards attainment of international benchmarks.

Acknowledgements

The authors are thankful to the anonymous referees for their valuable comments.

References
Basic Port Statistics of India (2011–2012) Transport Research Wing, Ministry of Shipping, Govt.
of India, New Delhi.
Bichou, K. and Gray, R. (2004) ‘A logistics and supply chain management approach to port
performance measurement’, Maritime Policy and Management, Vol. 31, No. 1, pp.47–67.
Bryan, J., Weisbrod, G. and Martland, C.D. (2007) ‘Rail freight as a means of reducing roadway
congestion: feasibility considerations for transportation planning’, Journal of the
Transportation Research Board, Vol. 2008, pp.75–83.
Dalkey, N. and Helmer, O. (1963) ‘An experimental application of the Delphi method to the use of
experts’, Management Science, Vol. 9, No. 3, pp.458–467.
De Langen, P.W. and Sharypova, K. (2013) ‘Intermodal connectivity as a port performance
indicator’, Research in Transportation Business and Management, Vol. 8, pp.97–102.
De Langen, P.W., Nijdam, M. and Van Der Horst, M.R. (2007) ‘New indicators to measure port
performance’, Journal of Maritime Research, Vol. 4, No. 1, pp.23–66.
Deshmukh, A. (2011) ‘Indian ports-the current scenario’, Working Paper No. 14, Dr. Vibhooti
Shukla Unit in Urban Economics & Regional Development, pp.1–22.
i-maritime (2003) India Port Report: Ten Years of Reforms and Challenges Ahead, i-maritime
Consultancy Pvt. Ltd., Mumbai [online] http://www.imaritime.com (accessed 14 December
2014).
Indian Port Association (IPA) (2014) [online] http://www.ipa.nic.in (accessed 3 December 2014).
International Chamber of Shipping (ICS) (2015) [online] http://www.ics-shippinp.org (accessed
2 July 2015).
Jung, B. (2011) ‘Economic contribution of ports to the local economies in Korea’, The Asian
Journal of Shipping and Logistics, Vol. 27, No. 1, pp.1–30.
 364 A. Mandal et al.

Kasypi, M. and Muhammad, Z.S. (2006) ‘A regression model for vessel turnaround time’, Tokyo
Academic, Industry & Cultural Integration Tour, December.
Lee, S.W. and Kim, C.H. (2006) ‘Performance evaluation of Asian port distriparks using factor
analysis’, Ocean Policy Research, Vol. 21, No. 1, pp.52–82.
Marlow, P.B. and Paixao Casacaa, A.C. (2003) ‘Measuring lean ports performance’, International
Journal of Transport Management, Vol. 1, No. 4, pp.189–202.
Ministry of Shipping (MoS) (2010) Performance Audit of Functioning of Major Port Trust in
India, Report No. 3 of 2009–2010, Chapter V, pp.1–15.
Ministry of Shipping (MoS) (2014) [online] http://www.shipping.nic.in (accessed 4 December
2014).
Puiga, M., Wooldridgeb, C. and Darbraa, R.M. (2014) ‘Identification and selection of
environmental performance indicators for sustainable port development’, Marine Pollution
Bulletin, Vol. 81, No. 1, pp.124–130.
Saaty, T.L. (1980) The Analytic Hierarchy Process, McGraw Hill International, New York.
Sanchez, R.J., Hoffmann, J., Micco, A., Pizzolitto, G.V., Sgut, M. and Wilmsmeier, G. (2003) ‘Port
efficiency and international trade: port efficiency as a determinant of maritime transport
costs’, Maritime Economics and Logistics, Vol. 5, No. 2, pp.199–218.
Trujillo, L. and Nombela, G. (1999) Privatisation and Regulation of the Seaport, Industry Policy
Working Paper 2181, World Bank, Washington DC.
UNCTAD (1976) Port Performance Indicators, United Nations, New York
UNCTAD (2013) Review of Maritime Transport, United Nations, Geneva.

View publication stats