You are on page 1of 37

Study on the Impact of Factors Characterizing the Port Performance

ISEG - School of Economics and Management


Technical University of Lisbon
Rua Miguel Lupi, 20
1249-078 Lisbon, Portugal
Phone: + 351 213 970 264, Fax: + 351 213 979 318
E-mail: cege@iseg.utl.pt;

Vitor Caldeirinha
ISEG - School of Economics and Management
Technical University of Lisbon
Rua Miguel Lupi, 20
1249-078 Lisbon, Portugal
Phone: + 351 213 970 264, Fax: + 351 213 979 318
E-mail: vitorcaldeirinha@gmail.com;

This paper may be submitted to the Journal of____________________.

February 2010

Abstract

This research work aimed to determine what is the impact of the characterization factors of
ports on performance. The study was conducted on a sample of 43 European ports, the
information was collected through surveys with a broad base of quantitative information
about the various characteristics of ports in terms of location, governance, size,
infrastructure, specialization, maritime services, logistical Integration, in terms of
characteristics of the port, and the performance of the region and operating and financial
performance and efficiency, as determined by Data Envelopment Analysis model. The
work consisted of testing a general model that seeks to explain the performance of the port
in its various forms based on the factors that characterize the port, operationalized through
a set of variables and quantified using the tool of regression. The application of the
instrument of the regression resulted in the confirmation of the importance of location,
governance, size, Infrastructure, Specialization, and Logistical Integration and Maritime
Services, as decisive for the operational and financial performance and efficiency of ports,
in different ways for each dimension of performance analysis.

Key-words: port performance, port characterization factors, efficiency


1.INTRODUCTION
The efficiency and performance of the ports are becoming increasingly important
worldwide. This phenomenon is related to the privatization of management of ports and
port terminals, joining as vital links of the logistics of global operating companies, which
leads to the greater importance of efficiency and performance of ports, for countries to
reach international competitive advantage (Tongzon, 2002).
Moreover, the containerization and intermodal transport are decisive changes in the ports in
recent years.The containerization has led to two major changes in the ports (Cullinane,
2005): (a) the globalization of service coverage, achieved through numerous alliances and
acquisitions (horizontal integration) in the liner industry, (b) provision of logistical services
in the extended international context, increasing the supply of owners door-to-door by sea
transport companies and offering added value in terms of the supply chain (vertical
integration).
In another case, the extension of transport infrastructures, the creation of large logistics
areas, ports and inland, interconnected and forming bipolar systems, and the increasing size
of ships with the selection to only the large-scale hub - ports , serving increasing hintelands,
led to greater competition between ports, where the hinterlands intersect more and more
(Wang and Cullinane, 2006).
Sea lines have a greater bargaining power, taking into account the wide offer of ports, and
more in a position to demand higher performance and efficiency, higher speed and quality
of service (Wang and Cullinane, 2006) and sometimes ports have to deal with the departure
of a container line to another port, as was the case of the port of Singapore.
Thus, the performance, efficiency and competitiveness of the ports are currently topics
among academic researchers and users of the ports, which demonstrates the relevance of
this study.
Measure and maximize the performance of the ports is now essential to the fulfillment of
port role in the logistics chain, in a context of increasing competition between ports and
increasing among logistics networks, in the hinterland and in the foreland.
Given that improvements in accessibility by land, the internationalization of economies and
the growth of intermodal ports allow more straightforward and more competitive in larger
hinterlands, increasing the power of choice of customer, the goal of ports changed to
increase your traffic beyond the normal economic growth (Haralambides, 2002 and
Notteboom, 2005), with the objective of increasing the maximum output for the same input
factors.
Being efficient is a necessity of modern container terminals in a competitive environment,
as this clearly has a strong impact on unit costs and then in the price and competitiveness
(Nottebom and Winkemans, 2001 and Robinson, 2002). But not only in the container
market, as well terminals for general cargo and bulk begin to enter in the race for
efficiency.
Europe is facing greater competition in the ports due to the nearness of their ports, than
most of the rest of the world and is therefore essential to study this phenomenon in the old
continent, so there is particular interest in studying European ports as a target population
this study.
Moreover, it is noted that the ports are a major determinant of the cost of shipping, with
emphasis on port efficiency, as verified Sanchez et al. (2003) in a sample of ports in Latin
America, demonstrating the importance of this factor to sea-based logistics.The cost of
shipping depends so heavily on the cost of ports and on the time of ships in port, since these
costs are usually proportional to the time in port.
In fact, if we understand that a port usually has the dual purpose of, first, to develop
commercially, generating more business value of their business cluster port and also
contribute to the development of the region, creating jobs and attracting investment,
business and industry to its proximity, working as a pole of development, any of these
objectives will certainly need to improve performance, either in terms of efficiency, and in
terms of operation.
To be able to achieve these objectives of the ports and modern logistics chains, it is
necessary to increase their performance and it is essential to try to understand the extent to
which intervention is possible in particular characteristics of the various ports, and it is
therefore necessary to identify the factors characterization of the port that determine its
performance and to determine its importance.

Recent theoretical approaches and Gaps


Chang and Lee (2007), made an extensive review of existing studies with regard to port
performance and inter-port competition and concluded that they are to study issues such as
what is the hinterland where the ports compete ?Privatization becomes even more
competitive ports? How to measure differences between the relative efficiency of ports in
competition? As the "hinterlands" are changing the face of restructuring of supply chains?
Studies on the ports can be divided into these major categories, according to Chang and Lee
(2007): selection of input ports, competitiveness policies, governance, ownership and
privatization, measures of efficiency, performance and productivity and cooperation,
alliances and acquisitions.
Studies of factors for selection of ports have establishment models of choice of ports for
ships and cargo: Slack (1985), Murphy and Daley (1994) and Buckman and Veldman
(2003), cited by Chang and Lee (2007), which attempted to analyze the weight of the
factors influencing the choice of ports.
With regard to competitiveness policies, Robinson (2002), refers to the need to analyze the
port rather than a single point, but as a part of the transport chain. Second, Mak and Tai
(2001) and McCalla (1999) cited by Chang and Lee (2007), who studied the best way to
build new ports and terminals to be competitive.
Regarding the study of cooperation, alliances and acquisitions, there are several examples,
according to Chang and Lee (2007), namely: Heaver et al., (2001), which refers to alliances
and cooperation that have taken place in ports worldwide and its influence on port
performance, Song, (2002), which shows that the capital structures of the ports affect the
strategies of cooperation and competition between ports; Yap and Lam (2006), who studied
the competition between ports in Asia during several years, and Christidis (2001), which
shows the trends of change in the transportation industry and globalization have
transformed the operating environment of ports and forced the ports to adopt cooperative
strategies and alliances globally to improve their efficiency and competitiveness.
According to these authors, with regard to measures of efficiency, performance and
productivity, have been many studies addressing this issue, and comparing the performance
of ports at different levels.Examples, Cullinane, (2005), Tongzon, (1995) and (2005),
Estache (2002), Song and Yeo (2004), and the DEA, "Data Envelopment Analysis", a
source of great use for many authors in recent years.
Finally, in its study of governance, ownership and privatization, many of them have been
carried out on the influence of trends in the world to include private companies in the ports.
Baird, (2002), studied in detail the processes of privatization that occurred in ports, Bird
(1963) and Rodrigue and Notteboom (2005), defined the model of development of ports in
its various forms of location, infrastructure and governance and ownership.Cullinane
(2002) studied the influence of ownership of the terminals in their performance, and Slack
and Frémont (2005) who distinguished between terminals operated by international
companies and terminals operated by national companies, assessing the influence of this
factor in the performance of ports.
In this context it is important to cluster sets of characterization factors, which have been
studied by researchers, and which affect the performance of ports, seeking to understand
their relationship, proximity or affinity, trying to determine what are the real factors that are
behind the groups factors and found what is its contribution to the performance of ports.
Cullinane et al. (2005) carried out the measure of the efficiency of a set of port terminals
and concluded that the determinants of efficiency are the location, the governance and
ownership of the port, indicating that in a second phase, it was necessary to apply a
regression model to try to explain the determinants of port inefficiency, which is considered
that it is this point a gap in the existing literature on the performance of ports.
In fact, it appears that many of the studies address the issue of port performance just
looking for new ways to measure and compare ports and port terminals, but it is not
explained the differences and why a port is more or less efficient than another, or why it has
better or worse performance.
In recent years researchers have focused mainly on measuring the efficiency of ports, and
why a port has a higher output for the same amount of inputs used. But it is also importante
to see what are the characteristics of the port's that influence efficiency.This is a "Gap" that
was identified in this study.
This is reinforced by Tovar and Trujillo (2007) that compared the efficiency of a wide
range of European ports and conclude that their work fails to explain the factors that
determine the different levels of port efficiency, which would be very important to help
improve efficiency and to be a real alternative to the road in Europe.
Estache et al. (2005) state that in addition to partial studies on productivity, the academic
literature on the subject of the ports is very limited and that there are essentially tests
studying the link between efficiency and ownership of ports and comparison rankings of
efficiency, failing to study the link between efficiency and other characteristics of the ports
at a broader level.
Moreover, the authors note that few studies focus on aggregate output variables, and almost
disaggregate by type of cargo, not identifying the other output variables, and few studies
rely on environmental control variables.
The port is in a estuary? Nearby is the city? It is near the Mediterranean? What is the GDP
per capita in the region where the port is inserted and that influence the characteristics of
the port and it performance? These are issues that should be evaluated.
One of the shortcomings of the literature is to understand the importance of each feature of
the port on performance, so you can improve the performance of ports.
In structural terms, for example, concludes Cullinane, (2002) that the size influences the
efficiency and Notteboom, Coeck and Broeck (2000), the location of the Hub terminal
allows greater efficiency then feeder ports. However, there are many factors that determine
the efficiency of ports, but few studies attempt to examine systematically all the sources of
change in efficiency based on the characteristics of ports.
Another gap that was detected by the author is the need of studies using the DEA (Data
envelopment Analisys) efficiency figures as output of the ports characteristics, except
Turner, Windle and Dresner (2004). In fact it seems to be a Gap important that there are no
studies that make use, for example, of the rate of port efficiency DEA as na output variable
in explanatory models of the factors that influence port performance.

Study Objective
The purpose of this study is to identify and analyze the impact of factors characterizing the
ports that determine its performance, both for the level of activity, and in terms of
efficiency and financial performance.
As a control model, is still considered the question: to what extent the performance of the
port is influenced by the performance of the region, while environmental factor, given that
this factor has a direct bearing on the factors that characterize the ports and indirectly on
their performance.
The aim of the study is to analyze the European ports, and for this purpose was used a
sample of the population of European ports, taking into account the great diversity in terms
of characteristics of ports, efficiency levels and in different regions where are located.
The author carry out this study is to ascertain the factors that characterize the port, which
have influence on their performance, which is essential for the construction of new ports for
the adaptation of existing ports and development of competitiveness policies ports and
terminals, in order to try to have a port industry more competitive.
The study is based on a sample of European ports with figures relating to constructs such as
size, governance, location, infrastructure, logistical integration, maritime services,
specialization and performance of the region where the port is situated, then proceeds to an
analysis of the multiple regression model with output of various performance indicators of
ports, including cargo handling for different types of financial performance of the port
authority and port efficiency, whose indices are calculated from the DEA methodology. It
aims to know the importance of each factor and sets of factors for various dimensions of
performance of ports.

2. LITERATURE REVISION
Multidimensionality Ports
The academic works that analyze the efficiency, productivity or performance of the ports
are scarce and according Tujillo Gonzalez (2007), the port size, location and governance
are key determinants of efficiency, which affect the capacity of the port itself, and the size
of the market in the hinterland of the port that can make use of that capacity and also
determine how this capacity is managed, and how the port interacts in the market, in
competition with other ports.
But the literature review of the ports is also scarce and have started by Estache et al. (2001),
Cullinane, (2002) and Wang et al. (2006), which gave a strong contribution.
The economic study of the port began in the '60s with the study of the structure of port
charges, skills and investment.The first studies on port efficiency have emerged in the 90s.
Recently, the efficiency and productivity have been major themes for port researchers,
since there have been major changes with the expansion and deepening of ports, with
improved technology, organizational change, privatization and specialization of ports input
and terminals , with impacts on efficiency and productivity that caused obvious diferences
between ports.
The studies on port efficiency can be classified into three major groups (Gonzalez and
Tujillo, 2007): The first includes studies with one-dimensional or partial indicators of
productivity of the port system and did not reflect the multidimensional reality of ports;
The second group of studies includes those with only a vision of the engineering side, using
terminal operational simulations and the queues theory; The third group, the most recent,
covers the frontier production using multivariate approaches in the inputs and outputs and
support the political and economic port decision.
The disadvantage of one-dimensional ports view, as it only compares a variable input with
a variable output, do not cover the special multidimensional and multivariate nature of
ports, which handle various types of cargo and have several inputs related to labor, capital,
land, among others.
This problem was only solved through the analysis of TFP (total factor produtivity), which
reflects the overall contribution of all factors relevant input and all outputs. Poitras,
Tongzon and Li (1996) made one of the first studies with the ports on the DEA model, Data
Envelopment Analysis, which reflects this multidimensional nature of the ports, analyzing
its performance.
In 90 years, new methodologies for measuring efficiency were used in studies of the ports,
but there was a lot of discussion about which method best defined the complex reality of
the ports.Studies have focused on the relationship between efficiency and reforms in the
ports, port ownership, size, transhipment, investment, hub ports (Noteboom et al., 2000)
and the efficiency and time (Cullinane et al., 2004).

Governance and Logistical Integration


The ownership and management of ports, namely governance, is considered one of the
characterization factors that influence port performance and efficiency (Liu, 1995), since it
believes that when a management or ownership of ports is on the public side, there is no
incentive to be held constant improvement of management efficiency of the ports, in
opposition to what occur in ports run by private companies which aim to profit.
The model of port management has evolved over the past few years, having been made
concessions for many port terminals around the world, with the spread of the traditional
model of Northern Europe, given their success in terms of performance.
For example, Estache, Gonzalez and Tujillo (2001), there efficiency gains from port
reforms in Mexico and using the methodology of the production frontier, demonstrated the
occurrence of gains 6 to 8% in the efficient use of infrastructure port with the concession to
operate the terminals.
At the end of the twentieth century, the vast majority of ports that were previously managed
by port authorities, have suffered various forms of privatization of their management, either
by granting long-term licensing of new projects or BOT, Build Operate and Transfer.
The weight of the terminals operated by private companies have grow and thus have
important influence on the performance of ports, contributing to their integration into new
supply chains and global operators of the terminals, adding new expertise used by terminal
operators in other terminals in the world and customer satisfaction and performance levels.
Barros and Athanassiou (2004) report that privatization is the best way to dramatically
increase the efficiency of ports and Lui et al. (2005) reported that the Chinese port
terminals with Sino-foreign partnerships (private) have higher levels of performance. This
latest study shows that management by private companies linked to international groups
enhances the performance of the terminals to allow the integration of innovative technical
knowledge developed within the group and raise standards of performance through the
comparison in the group, or in order to meet the standards customers of the group.

Port Size
The size of the port has been, for many years, considered another factor that influence port
performance (Liu, 1995, Wingmans, 2003), one of the key variables taking into account
that affects economies of scale and agglomeration.
The productivity of ports increases with the size and there are significant economies of
scale, which led to a recommendation to invest more in bigger ports and be cautious in
small ports, (De Neufville and Tsunokawa, 1981).
However, small ports have also played a part, with economic impact in the region, despite
its poor performance in absolute terms.
It is interesting to know if a port can become larger, with better performance, if increase the
investment.The question usually arises more on the terminals than the ports, and yet the
terminals at larger ports may also benefit from this in its performance, in the result of
synergies with other terminals.
In 2005 Estache et al. decomposed the change in the efficiency of technical changes and
changes of scale, showing the importance of the performance of ports (like Turner, Windle
and Dresner, 2004, Gonzalez and Trujillo, 2007), checking the learning effect of the larger
ports that contributes to its better performance.
The learning effect is cited by many authors as an explanation for the difference in
performance between large ports and small ports, because larger ports are required to adopt
systems and processes to achieve more efficient to move a large cargo amounts, to be more
efficient and productive. The labor force of larger ports have access to more training result
too
In large ports, the problems occur more frequently, solutions are rebuilt several times,
authorities have created greater coordination and the systems are optimized.
The effect of scale and dilution of indirect costs, and fixed administrative costs are usually
indicated as the main contributors to the effect of size on the performance of ports.
Ports have, from time to time, to invest large amounts in infrastructure, which can lead to
some periods of inefficient, aiming to increase their size and have a better performance in
the future.

Location of Port and Region Performance


The location of the port seems to be another factor of Port performance (Lui, 1995) and
perhaps the most important, since the port does not exist by itself, with the exception of
ports exclusively "transhipment", but is dependent on development and performance of its
hinterland, despite the fact that the improvement of accessibility and land development of
railways have extended the hinterland of the ports to large distances, increasing competition
between ports, the basis for development of a port remains in its its closer hinterland.
In a study on the competitiveness of Chinese ports, using the Analytic Hierarchy Process
methodology and using a wide range of factors, Song and Yeo (2004) reported that the
volume of cargo from ports is strongly associated with the location, which can not be
changed regularly but location is a variable that, once built the port, can not easily be
changed, other than with the relocation of the port itself or specific terminals in order to
take advantage of any location.
In 2005, Rodrigue Notteboom and identified a new phase in the life of the ports in general,
they call regionalization, stressing the importance of the relationship between the
development of the port and development of the region where the port is located.
Ports have a pattern of change over time, appearing usually associated with a city, very
close to it, but gradually, with growth, going more and more away from the town in order
to obtain larger areas, deeper depth and avoid congestion on land accessibility and land
conflicts with their areas of expansion.
The very concept of location of the port is associated with various dimensions such as
distance to urban area, population density and richness of the region of influence, the
physical location on the coast, in a river or estuary, the location and accessibility in the face
of the industry, the location in a island, the mainland or in the peripheral zone, its distance
from the center of Europe, the cultural influences on organizational models, the situation
facing the major shipping routes and inland. No doubt the performance of the region
influence the characteristics of the port, except that the location factor contain elements of
the region.
The importance of the hinterlands of ports as their own extensions was also analyzed by
Guthed (2005), determining its performance.
Tongzon stated in 2002, the location is one of the main reasons for choosing the port of
Bangkok, which increases its operational performance, the location near small economies
affects performance of the port and points out that the demand for port services derive from
the size of cargo flow and consumption in the region where the port is located.

Infrastructure and Accessibility


The investment in port infrastructure and capital intensity in the ports has often been an
explanatory factor for differences in performance and efficiency in ports (Liu, 1995), since,
in fact, without the infrastructure and supply capacity, would not be possible to have a
higher movement of vessels or cargo.A large movement of vessels is only possible with
quay and equipment sufficient to allow not have high waiting times, unbearable for ships.
Moreover, a high level of efficiency in the port, which would provide a competitive
position in the port sector, need adequate infrastructure and superstructure exploited
intensively, to ensure the use of investment with high standards of performance.
In 1996, S. Achish concluded that capital investment was a major factor affecting the
productivity of ports in Israel. Founded that the extent of activity and capital investment
were the main influences on productivity. Not only should interest the amount of capital
invested, but also the quality of these investments, training and suitability of such
investments to market needs and demand.
Moreover, Goss (1990) states that the competition can lead to increased efficiency, but also
can lead to excessive investment in capacity of the port infrastructure.Although excess port
capacity is essential to maintain competition between ports, thereby maintaining its
performance in the customer point of view.
Sachish and Kim (1986), assessed the impact of labor and capital investment in the
performance of the port of Ashdod (Israel), and also found a relationship between
investment in container technology and improved performance.
The capacity of the quay is a variable input important for the efficiency that has been
studied by Park and R.K. P. De (2004), as input which is related to the results output.
In 2007, Garcia-Alonso and Martin-Bofarull, in a study on the evolution of the relative
efficiency and the hinterlands of the ports of Valencia and Bilbao, during a period of heavy
investment in both infrastructure, stands that there not always the same level of investment
in infrastructure leads to similar improvements in performance, it is necessary to study
other factors related to the location, integration into supply chains, the hinterlands, among
others.

Shipping Services and Logistical Integration


In 2002, Tongzon, studied the determinants of port performance and the ports choice to
determine that port efficiency is the most important factor in choosing a port, and has
addressed the equipment, the frequency of ships, infrastructure, location, rates and
productivity indicators, transit time and waiting time for ships in port. In its 1995 study,
reported the frequency of ships of the line and routes that pass through the port as important
factors in choice of port and its performance as well as the importance of economies of
scale of larger vessels to carry port.
Furthermore, the lines determine the ports that call based on the partnerships that have and
logistics networks that integrate (Tongzon and Heng, 2005), and the important issue of
integration of ports and maritime services, including the links of global operators to major
ports worldwide and the level of integration in the logistics networks.
In 2003, Veldmen Buckmann to explain the market share of the northern ports of Europe
and its performance using factors such as frequency and transit time of vessels and freight
prices, the prices of the terminal and inland transport.Turner, Windle and Dresner (200 4)
studied the impact of the type of shipping services and port facilities in the performance of
ports.
The frequency of liners and size of vessels calling at the port are very important
determinants for the efficiency of the port itself, since that characterize the service that the
port provides to its customers and cargo, and the value it adds for customers.
A liner requires pre-determined schedules, ports of origin and destination, freight pre-
defined integration with chains of land and sea transportation, higher number of weekly
lines is important to attract more cargo at to thet port, increasing the performance of the
port, minimizing wait times and costs for transportation sea, offering a wider range of
destinations to lower costs and with low "transit times".
A port with regular lines has usually better performance and higher levels of efficiency and
is required to maintain these levels to keep the lines and attract new regular lines that have
a very high requirement for quality of service and schedules .
Basically, it can be said that the maritime services and the integration of the ports in global
maritime logistics, or even regional levels allows better performance to ports, making it
more attractive.

Specialization
Specialization, including the level of containerization, are reported by Tovar and Trujillo,
2007, Medda and Carbonaro, 2007 and Laxe, 2005, but no less important is the rate of
unitization cargo (general cargo/ total cargo of the port), as reflecting the degree of
development of the port, to the phase for modern industrial and commercial port.
In a study of Caldeirinha, 2007, it was found that the Iberian ports are divided into three
quadrants according to the crossing of two variables, the rate of unitization of cargo and its
composition, which explains a good part of their characteristics and thus the performance.
In fact, the ports with greater specialization in containers usually have higher levels of
income per ton, and often more efficient use of their infrastructure platform. A port
specialized in bulk has usually a very high performance in number of tons, but low
performance in terms of revenue per ton and per employee.
A not specialized port, do not usually have high levels of efficiency, because it
infrastructure is flexible and suitable for all types of cargo, not getting the most out of each
type of load, because there are no specialized equipment and there is a constant adaptation
of the equipment to each type of cargo. This ports are usually more expensive.
Specialization is thus an important variable to consider in relation to the performance of
ports, which may serve to classify groups of ports.

Performance of ports
As previously stated, the performance of ports is now essential for all stakeholders in its
operations, including managers, customers, traders, industries and governments.
But the performance can be measured in various ways, may be a synonym for efficiency,
doing more with less. A port may only do more in absolute terms, irrespective of the
resources spent up, but others may prefer the performance level of costs, guaranteeing a
minimum service.For others it may be possible to earn more for each ton moved, providing
more services.
So consider the authors studied a range of performance indicators that vary according to the
purpose they want to, or the target audience they have in mind, or the phenomenon in
analysis.
The performance indicator most used is the absolute cargo in tons or TEUs (twenty-foot
equivalent unit) in the case of containers. Some authors use the distinction between
absolute tonnes by different type of cargo and number of ships served by the port in a given
period. This indicator shows the breadth of choice of port for their customers, ie, more
cargo, shows that more customers have chose the port and shows better operational
performance.In this case, each ton has one vote in the choice of port.
Some prefer to use indicators of absolute income of the port authorities or per tonne of
cargo. The cost of the port to the client, related to the revenue of the port, is sometimes
used as a factor that characterizes the port and as a choice factor of the port. In fact, this is
an important factor in choice of port by cargo and vessels, but is also an indicator of
performance resulting from the features of the port itself, so it is vital understand which are
the determinants of this variable.
Now, with the greatest concern for the efficiency of ports, the efficiency indicator
multidimensional DEA has been widely used in the comparison between ports, although
few studies have used this indicator in regression models which can explain the factors that
determine the values found in the ports. It should be noted that the efficiency is considered
by many authors as one of the factors influencing the choice of the a port.
How to choose the performance indicators of the port? We must choose those related to its
internal efficiency, with its size and market success? These indicators are certainly related,
but are sometimes conflicting.
Ships do not choose the ports with the worst performance in terms of waiting times and
speed of operation, due to the time of detention of ships, because that have a significant
cost in shipping and it must be added to price charged by the port to ships.
And in the case of general cargo liner market, cargo often follow the choices of ships,
which is not the case in the tramp market of bulk, where the ships follow the preferences of
the cargos in terms of ports. What audience to choose, as this influences the indicators to
select?
If we consider how the performance qualities that lead to attract more ships and more
cargo, we have for example Sanchez et al. (2003) determined the three main components of
port efficiency: the time in port, efficiency in the terminal and the turnaround time of
vessels in port.
But the main performance indicators used in ports in almost studies are the movement of
cargo, either in tons or TEU, or by type of cargo, roll-on roll-off, breakbulk, containerized
cargo, solid bulk and liquid bulk, since this is the end result of any port, move more and
more cargo and ships.
Just to name a few authors who used the throughput of the ports in absolute and variable
output model of performance analysis, refer to Song and Yeo, 2004, Poitras, Tongzon and
Li, 1996 Barros, 2003 and Trujillo Tovar, 2007 Garcia -Alonso and Martin-Bofarull, 2007
Park and De, 2004 Herrera and Pang, 2006.
Another indicator of performance that seems appropriate is the level of revenue per tonne
or employee of the Port Authority in, the public perspective, since it reflects the output
value of the range of services offered and what customers are willing to pay in terms of
rates to call the port, given its condition or location.Just to name a few authors who choose
the performance indicators of revenue we have Barros, 2003, Park and De, 2004, Kent and
Ashar, 2001, Gonzalez and Trujillo, 200 7, Turner et al, 2004.
Basically, the performance of a port is a multivariable reality, and in this study used a
battery of performance indicators that go through the operational performance in the
handling of cargo and a year, the financial performance of port authorities and the
economic performance or efficiency the port in its relationship between output and inputs,
in terms of DEA indicator.

3. THE CONCEPTUAL MODEL


Research model
The test model is based on the hypothesis of the relationship between the characteristics of
ports and their performance. It is considered that the performance of ports is explained
largely by the characteristics of ports, which determine its level. In other words, it is
considered that the ports with different characteristics will have different performance
related with this characteristics, in the various levels at which performance can be
measured.
The performance of the region also influences the performance of the port directly or
indirectly, through its influance the characteristics of ports. In this model, the performance
of the region in which the port is located and to which it relates serve as an environment
variable or control.
The explanatory model of port performance based on their characteristics makes use of
constructs already developed, setting variables that act as drivers.
Basically, we identify, in the literature review, a set of constructs based on factors that
characterize the ports that have emerged as influencing the performance of ports.
To characterize the ports were considered seven constructs which appear to us as somehow
separate and characterize the ports, seems to contribute to their performance in a relevant
way.
In terms of performance of ports, we used three constructs covering multidiversidade port
performance, taking into account the objectives of port in operational and financial
efficiency.

Figure 1
Model

Region Performance

Location

Size Operational Performance

Infrastructure

Specialization Port Caracteristics Port Performance Financial Performance

Maritime Services

Logistical Integration Port Eficiency

Governance

The constructs of the input model considered for the characteristics of the ports are the
location of the port, the port size, port infrastructure, port specialization, shipping services,
the degree of integration into global logistics and governance model.
The constructs of the output model considered for the performance of the port, are
operating performance, financial performance and performance in terms of efficiency of the
port.
It is also important to better explain what are the variables drivers considered as proxies of
the constructs to operationalize the model.
To simplify the processing of data was defined a simple classification for each variable:

Table 1
2008 variables Construct
DROTERD2 Distance to Rotterdam in a straight line LOCATION
in km
DMEDIT3 Distance from to the Mediterranean Sea LOCATION
axis east-west in km
SEAPORT4 Sea Port (1) or River / Estuary (0) LOCATION
DCITY5 Distance to nearest city in Km LOCATION
QUAYL6 Total length of quay in meters SIZE
CRAINSKM7 Number of cranes / km of quays INFRASTRUCTURE
TERMSIZE8 Average size if terminals in tons INFRASTRUCTURE
MAXDRAFT9 Maximum depth of quay INFRASTRUCTURE
TXUNIT10 Unitization rate (general cargo in SPECIALIZATION
tons/total cargo in tons)
TXHORIZ11 Horizontalization rate (roro cargo in SPECIALIZATION
tons/general cargo in tons)
TXCONT12 Containerization rate (container cargo in SPECIALIZATION
tons/general cargo in tons)
REGULARLSHIPS13 Number of regular lines / total number of MARITIME SERVICES
ship calls
SHIPSIZE14 Average Ship size in tons of Gross MARITIME SERVICES
Tonnage
BIGSHIPO15 Number of regular lines of the seven big LOGISTICAL INTEGRATION
ship operators / Total number of regular
lines
PORTPRIV16 Percentage of private terminals in the GOVERNANCE
port: (1)> = 50%, (0) <50%
GDPCAP17 GDP per capita of the region / GDP per REGION PERFORMANCE
capita of the EU27
TOTALTON18 Total cargo in tons OPERATIONAL PERFORMANCE
GENERALTON19 General Cargo in tons OPERATIONAL PERFORMANCE
BULKTON20 Bulk in tons OPERATIONAL PERFORMANCE
EURPERSON21 Port Authority Income per employee FINANCIAL PERFOMANCE
EURTON22 Port Authority Income per Ton FINANCIAL PERFOMANCE
DEABCC23 DEA* BCC efficiency indicator EFICIENCY
DEACCR24 DEA* CCR efficiency indicator EFICIENCY
*DEA – Data Envelopment Analysis

Variables and constructs


The location influences the performance of the port. A better location lead to better
performance at the port, facilitating its expansion, its access, its construction, its operation,
its characteristics or its market and its shipping services.
The location refers to where the port is implanted, in physical terms, whether it is an
estuary, a river, the sea, which can constrain or facilitate the development of the port and its
performance in terms of ships access, design of the infrastructure and rail access.
The location also refers to the distance between the port and the city or the nearest urban
area, which may restrict the expansion of port or congest the land access.
The location also refers to the distance to major roads or points of cargo handling.In
Europe, in the main references in this field are the Port of Rotterdam, logistics center in
Europe, the most populous region, and the Mediterranean Sea, through which the lines'
round-the-world take the mother-ships.
The size of the port is an important variable in performance. The ports have larger
economies of scale, creating a greater attraction for cargo, that effects fixed costs, the
number of maritime services to more destinations and more regularly by the effect of
learning and the organization system more flexible and results-oriented.
The size of a port is related to its capacity, ie the number of ships that the port can receive
simultaneously, which leads to the variable length of quay, as the best to emulate the scale.
The characteristics of infrastructure are important to understand the performance of a port.
The ports with the highest intensity of handling equipment for vertical loads per meter of
quay can perform better, since dispatch ships faster, and can apply more than one crane per
vessel, reducing the cost of stay, maximizing the quay use and the return on heavy
investments.
The ports with large terminals may have important effects on it performance the result of
economies of scale of these terminals and their learning effect and return on fixed
investments and overall management costs, facilitating the logistics between lines, the
transhipment and the relations of interface with the hinterland, creating benefits for clients,
including logistics and value added to the cargo, without further handling operations.
The ports with the largest depth access may receive larger vessels, generating positive
effects on the performance of the crane and the terminal itself, reducing the time of ships
maneuver by reducing freight costs in the port and increasing the productivity of
operations.
The ports can be classified as ports specialized in certain types of cargos.The kind of
specialization of the port determines its physical characteristics, its infrastructure, its traffic,
vessels, helping to explain differences in performance between some ports.
A port with a higher rate of unitization, ie, with a greater movement of general cargo in
tons thenbulk cargo, tends to be a port with better results in terms of revenue per tonne, has
more added value in the port.A bulk specialized port tends to be a port with higher
productivity, since the spacialized systems in bulk get higher speeds, because the
homogeneity of cargo and discharge in a continuous process.
A port specialized in the horizontal cargo, roll-on roll-off, cars and heavy vehicles, trailers,
among others, tend to have lower berth productivity due to the nature of the operation,
although operations do not require investment in equipment.
The importance of regular lines in the port and the size of vessels calling the port are very
important determinants for the efficiency of the port itself, since that characterize the
service that the port provides to its customers and cargo, as well as adds value to customers.
It is not enough to have private management at the port, the question is what is the degree
of integration of private companies that operate the port in the regional and global logistics,
since this may affect the efficiency of the port.
In fact, some ports have been privatized under the operation of local small companies that
do not allow raising its performance globally, keeping roughly the same rates of growth.
Since the ports and terminals that have been licensed to major international operators,
including global ship operators, seem more prone to develop their performance, as they are
integrated into strings of international ports, drinking their technical knowledge and being
"pulled" by international group to levels of much higher efficiency, and integrate group's
international supply chains.
The integration of ports in chains allows performance much higher than the ports that do
not have it, and that is a factor which characterizes the port and its services.
The model of governance may be related to the performance of ports, it is important to
study additional evidence, seeking to identify the benefits of private management of ports
and what is the public and private mix that maximizes the output port.
In fact, this is one of the main variables studied by several authors who address the issue of
ports, comparing the privatization impact on ports performance.
The performance of the region has a significant impact on the performance of the port. In
fact, the region determines its infrastructure and facilities, and determines it hinterland and
perfomance. A larger performance of the region, determine a port with higher performance
than others.
Performance
The performance is the result of the economic operation and the existence of the port,
which can be approached from different perspectives, depending on your point of view. We
will consider the performance operationally, financially and in terms of relative efficiency.
The level of annual turnover of the port is commonly used by researchers as an indicator of
operating performance of the port can be introduced in studies such as the total movement
of the port where tons of cargo, or more particularly in TEU.
However, it do not compare ports that do not have any containers, so it may not be helpful,
except for comparison exclusive of specialized container terminals.
The basic variables of a port such as the movement of goods total (tonnes) and the
Movement decomposed (general cargo and bulk) are used as variables to the performance.
In terms of financial performance, the variables total revenue for the Port Authority euro
per tonne and total revenue of the Port Authority per employee are used reflecting the
added value achieved by the port with its activity as continuous variables monetary greater
than zero.
This type of variable is used by Ugboma, 2006, Poitras, 2005, Kent, 2001; Tongzon, 2005,
2002, Lee, 2006 and is defined as the total revenue of the Port Authority (PA) in the year,
on the total number of tons of cargo moved on the year, in euros, and total revenue of the
Port Authority (PA) with the port in the year, the total number of employees in euros.
Finally in the case of performance in terms of efficiency, we use two variables related to
the Index of Relative Efficiency DEA, one for the model of constant returns to scale (CCR)
and one for the model of increasing returns to scale (BCC).
Both variables are calculated based on the use of two output variables - tonnes of general
cargo and tonnes of bulk - and the three input variables - the sum of the lengths in meters of
quay, the sum of area of terminals in square meters and total number of cranes in the port.
This type of variables is used by Ugboma (2006) and Turner (2004) as output variable in
regression models of performance of ports.
The efficiency measure consist in to compare the level of input of productive factors of
different nature for the production of one unit of output port.
For example, Liu et al., 2005, defined as DEA model inputs to ports, the land factors (based
on the variable size of the area), capital (based on variable length of quays, emulator of
investment in port) and work (based on variable number of cranes, which is proportional to
the number of workers).
These indicators are described in the literature review of Gonzalez and Trujillo, 2007, and
are frequently used by many researchers. The novelty is to use this s indicators as output of
a regression model for ports.

4. RESEARCH METHODS
Data
According to the 2007 annual report of ESPO, European Sea Ports Organization, Europe
has 350 ports according to information from Eurostat, and from those 230 ports have more
than 500,000 tons of bulk annually and 200 ports have over 200,000 tons of general cargo.
The list of ports highlighted in the annual report for 2007, ESPO, was crossed with the
information databases of emails from port authorities contained in the following sites:
http://portfocus.com/index.html and http://www.worldportsource.com/countries.php.
Then was sent the e-mail surveys of the port authorities of 230 ports, and during the months
of April to June 2009 were received 43 complete responses to the survey.
Methodologies

DEA - Data Envelopment Analysis


DEA (Data Envelopment Analysis) was originally developed by Charnes, Cooper and
Rhodes (1978), and allows to compare the relative efficiency of complex production units,
such as ports, schools, hospitals, banks, among others.
Since its first publication in 1978, the DEA has developed dramatically, and that many
applications have shown how this technique can be considered an important tool for
assessing efficiency.
Ahn, Charnes and Cooper (1988) evaluated the relative efficiency of public and private
universities. Morey, Fine and Loree (1989) used DEA to determine the relative
performance of public hospitals in California.In banking sectors, DEA has been used to
identify efficient units (Di Giokas, 1990).
This methodology measures the relative efficiency of units of decision making (DMU-
Decision Making Units), which perform tasks with multiple inputs and multiple outputs.
The DEA involves the task of selecting inputs and outputs to produce an empirical
production function that is based on optimal behavior observed.The DEA model compares
each DMU with the best practice observed, to obtain a measure of relative efficiency. Each
DMU is then classified as efficient or inefficient (Moita, 1995).
The DEA has been structured along two different models, known as CCR developed by
Charnes, Cooper and Rhodes and BCC developed by Banker, Charnes and Cooper.
The CCR, which was used in this study is to evaluate the overall technical efficiency, but
admits the possibility of constant returns to scale, that is, if a unit assessed increase
resources at a certain level, its production should increase at the same proportion, as if this
unit to reduce the resources, their production should be reduced proportionately. The BCC
model allows increasing returns to scale, and seems to apply to the ports, due to economies
of scale and learning effect.
DEA analysis is a technique of Operational Research, which is based on linear
programming, and aims to analyze comparatively independent units.
Because it is a nonparametric tool, DEA differs from parametric approaches, and optimizes
each individual observation in order to calculate an efficient frontier, determined by units
that are Pareto efficient.
This technique has been applied in several studies relating to transportation systems.
Novaes (1997 and 2001) and Chu and Friefding (1992) and Odeck And Halmarsson Jr.
(1996). In the port sector, studies developed by Bendall and Stent (1987) Tabernacle
(1995), Ashar (1997) and De Monie (1987). According to Cullinane et al (2004), the DEA
is one of the most important techniques to measure efficiency. The author also stresses that
there are many applications of DEA in the classic seaport industry like the examples of
Tongzon (2001), Valentine and Gray (2001) and Martinez et al (1999).
As mentioned above, the DEA technique to classical two angles of analysis, the CCR
model which determines a border CRS (constant returns to scale) indicating that growth of
inputs will produce proportional increases proportional outputs.
The BCC model determines a border VRS (Variable Returns to Scale) and differs from the
CRS model by considering the possibility of increasing returns or decreasing returns to
scale on the efficient frontier.
It is assumed that there are n decision making units (DMUs) to be evaluated. Each DMU
consumes varying amounts of m different inputs to produce s different products.
Specifically, DMUj (j = 1 ,..., n) consumes an amount Xj = (xij) of inputs (i = 1 ,..., m) and
produces an amount Yj = (Yrs) product (r = 1 ,..., s).
It is assumed that xij> 0 and Yrj> 0.
The array of products sxn is represented by Y and the mxn matrix of inputs is denoted by
X.

Figure 2 – CCR Models

Badin (1997)

Figure 3 – BCC Models

Badin, 1998

In this study, the CCR and BCC models, maximizing production, are based on the Frontier
Analyst software version 4.

Linear Regression
"The analysis using regression models is used as a statistical tool that seeks to find the
relationship between two or more variables so that a variable can be calculated from one
another or others."(Neter & Wasserman, 1985).
This form of a relationship through a regression model is different from that found by a
causal role, as a function and has a perfect relationship between the variables, the
relationship found by regression models is not exactly perfect, and have distortions in the
estimated parameters.
The purpose of a regression model is to determine a relationship between the data so that a
variable can be defined in relation to another or others.This relationship is not perfect as a
function has errors in the estimated values, called errors of dispersion.
There are several regression models, and the choice of a model depends on the
characteristics of the data and the objective to achieve with the regression.
According Neter & Wasserman, 1985, the regression model is a methodology that seeks to
show two essential ingredients of the statistical relationship.On the one hand, the trend of
the dependent variable Y vary with one or more independent variables X, a system, and
second, show the set of observations around the curve of the statistical determined.
5. DATA ANALISYS
Descriptive Statistics
In the following we find the main descriptive statistics of the variables that are used in the
model.

Table 2
Descriptive Statistics
N Minim um Maximum Sum Mean Std. Deviation Variance Skewnes s Kurtos is
Statis tic Statis tic Statis tic Statistic Statis tic Statistic Statis tic Statistic Statis tic
DROTERD2 43 80 2970 49392 1148,65 670,777 449941,42 0,246 -0,45
DMEDIT3 43 0 3120 52741 1226,53 848,101 719275,26 0,345 -0,816
SEAPORT4 43 0 1 31 0,72 0,454 0,206 -1,021 -1,006
DCITY5 43 0 32 154 3,59 5,865 34,395 3,385 13,744
QUAYL6 43 400 80000 365938 8510,19 13246,584 1,76E+08 4,193 20,815
CRAINSKM7 43 0 12,7 156,001 3,62793 3,033383 9,201 1,231 1,09
TERMSIZE8 43 86765 1,24E+07 1,07E+08 2,48E+06 2,72E+06 7,41E+12 2,177 5,039
MAXDRAFT9 43 7 26 590 13,73 4,525 20,475 0,823 0,801
TXUNIT10 43 0,02528 1 18,86606 0,4387456 0,30757091 0,095 0,442 -1,093
TXHORIZ11 43 0 0,98731 9,09449 0,2114998 0,29902372 0,089 1,58 1,521
TXCONT12 43 0 0,98116 15,14615 0,352236 0,31767226 0,101 0,514 -0,901
REGULARLSHIPS13 43 0 0,01408 0,1686 0,0039209 0,0033917 0 1,16 0,837
SHIPSIZE14 43 59 20026 243189 5655,55814 5,33E+03 2,84E+07 1,266 1,092
BIGSHIPO15 43 0 1 7,4344432 0,17289403 0,280520875 0,079 1,742 2,397
PORTPRIV16 43 0 1 24 0,56 0,502 0,252 -0,243 -2,038
GDPCAP17 43 50 127 3872 90,05 25,548 652,712 -0,308 -1,118
TOTALTON18 43 209000 178675809 1001260463 23285127,1 3,61E+07 1,30E+15 2,977 9,644
GENERALTON19 43 145000 117322134 421787594 9809013,81 2,27E+07 5,16E+14 4,105 16,951
BULKTON20 43 0 81051000 578971810 13464460,7 1,84E+07 3,38E+14 1,955 3,937
EURPERSON21 43 17,8 931 7733,1 179,84 165,4261 27365,78 2,962 10,827
EURTON22 43 0,0005 9,6 126,3005 2,937221 2,1714463 4,715 1,437 1,932
DEABCC23 43 3,5 100 2115,5 49,198 36,8929 1361,087 0,417 -1,53
DEACCR24 43 2,4 100 1311,9 30,509 31,8945 1017,258 1,442 0,751

The normality of the variables used in the model was tested with the Kolmogorov-Smirnov
and Shapiro-Wilk, and found that only not rejects the hypothesis H0, hypothesis that the
variable distribution is normal for the parameters of average and the variance, for the
variables DROTERD2, DMEDIT3, MAXDRAFT9, TXUNIT10, TXCONT12 and
SHIPSIZE14.This happens due to small sample size.
Thus, it should be some caution in the analysis of model results, given this constraint,
which may have significant effects on the power of parametric tests, and for non-parametric
(Maroc, 2007).

Correlation
An important stage prior to the application of variables to the model is the correlation of
data, for which, given the large number of variables involved, we used the Pearson
correlation between variables. Note that all the statistical calculations of this study were
performed with the use of SPSS software, version 1 7.
In the analysis of the contents of the Pearson correlation between variables, it was possible
to draw some important conclusions. The correlations between factors and dependent
variables are mostly significant at the 1% to at least one of the dependent variables, except
for the variables DMEDIT3, DCITY5, TERMSIZE8, BIGSHIPO15 and PORTPRV16,
although this not mean that these variables, combined with others, do not contribute
significantly to the explanation of the evolution of some dependent variables.
There is significant correlation between the dependent variables TOTALTON18,
GENERALTON19, BULKTON20 and EURTON22, and there is significant correlation
between the dependent variables DEABCC23 and DEACCR24, which seems normal
considering that these are performance variables resulting from the operation of the port or
its efficiency .
Just outside is the dependent variable EURPERSON21, since it is a variable that is also
dependent on the efficiency of the port authority.
The variables used as inputs have several intense relations between them, which lead to
concern with issues of multicollinearity when the regression model, which is the objective
of this study.

Table 3
Pearson Correlations
REGULARLSHIPS13

GENERALTON19

EURPERSON21
TOTALTON18
PORTPRIV16
MAXDRAFT9
CRAINSKM7

BIGSHIPO15

BULKTON20
SHIPSIZE14
DROTERD2

TERMSIZE8

TXHORIZ11

DEACCR24
GDPCAP17

EURTON22

DEABCC23
SEAPORT4

TXCONT12
TXUNIT10
DMEDIT3

QUAYL6
DCITY5

DROTERD2 1 -,372* 0,079 -0,01 -,335* -0 -0,1 0,023 0,108 -0,25 0,053 0,117 -0,09 0,019 -,391** -,464** -,420** -,379* -,354* -0,03 -0,07 -0,12 -0,13

DMEDIT3 -,372* 1 -0,09 -0,01 0,126 0,09 0,192 -,363* 0,214 ,306* -0,26 -,311* -0,24 0,043 -0,04 0,274 0,043 0,129 -0,08 0,185 0,256 -0,05 -0,04
SEAPORT4 0,079 -0,09 1 -0,1 -,351* -,321* 0,066 0,232 -0,08 0,256 -0,07 -0,11 0,048 0,084 -0,24 0,044 -,320* -,335* -0,22 0,283 0,149 -0,13 0,006
DCITY5 -0,01 -0,01 -0,1 1 -0,01 0,072 -0,11 0,283 -0,21 0,004 -0,04 -0,01 0,257 0,031 0,238 0,031 0,083 -0,05 0,22 0,273 -0,23 0,093 -0,03
QUAYL6 -,335* 0,126 -,351* -0,01 1 0,004 0,121 0,23 0,088 -0,11 ,328* ,312* 0,219 -0,01 ,309* ,303* ,829** ,889** ,528** 0,062 -0,21 0,18 -0,04
CRAINSKM7 -0 0,09 -,321* 0,072 0,004 1 0,217 0,076 -,353* -,420** 0,277 0,136 0,177 ,352* 0,019 -,385* 0,173 0,069 0,254 -0,14 -,342* -0,06 -0,19
TERMSIZE8 -0,1 0,192 0,066 -0,11 0,121 0,217 1 0,021 -0,03 0,11 0,063 -0,17 0,089 0,176 0,196 -0,06 0,164 0,131 0,161 -0,06 -0,03 0,076 0,232
MAXDRAFT9 0,023 -,363* 0,232 0,283 0,23 0,076 0,021 1 -,430** -0,24 ,391** 0,264 ,703** 0,17 ,364* 0,069 ,319* 0,18 ,403** ,384* -0,3 -0,05 -0,06
TXUNIT10 0,108 0,214 -0,08 -0,21 0,088 -,353* -0,03 -,430** 1 ,544** -0,2 -0,07 -,580** -,435** -0,15 0,198 -0,04 0,23 -,358* -0,05 0,258 -0,05 -0,1
TXHORIZ11 -0,25 ,306* 0,256 0,004 -0,11 -,420** 0,11 -0,24 ,544** 1 -0,3 -,315* -,373* -0,22 0,008 ,371* -0,12 -0,01 -0,22 0,119 ,533** 0,018 0,023
TXCONT12 0,053 -0,26 -0,07 -0,04 ,328* 0,277 0,063 ,391** -0,2 -0,3 1 ,619** ,321* 0,285 0,109 0,04 ,435** ,375* ,391** -0,05 -0,23 0,039 -0,12
REGULARLSHIPS13 0,117 -,311* -0,11 -0,01 ,312* 0,136 -0,17 0,264 -0,07 -,315* ,619** 1 ,312* -0,05 0,015 0,239 ,399** ,384* ,309* 0,002 -0,28 0,156 0,031
SHIPSIZE14 -0,09 -0,24 0,048 0,257 0,219 0,177 0,089 ,703** -,580** -,373* ,321* ,312* 1 0,182 0,208 0,025 ,438** 0,217 ,591** ,403** -,468** -0,01 -0,02
BIGSHIPO15 0,019 0,043 0,084 0,031 -0,01 ,352* 0,176 0,17 -,435** -0,22 0,285 -0,05 0,182 1 -0,16 -0,28 0,062 -0,05 0,186 -0,04 -0,21 0,184 0,142
PORTPRIV16 -,391** -0,04 -0,24 0,238 ,309* 0,019 0,196 ,364* -0,15 0,008 0,109 0,015 0,208 -0,16 1 ,328* 0,287 0,256 0,247 -0,04 -0,06 0,13 0,25
GDPCAP17 -,464** 0,274 0,044 0,031 ,303* -,385* -0,06 0,069 0,198 ,371* 0,04 0,239 0,025 -0,28 ,328* 1 ,313* ,376* 0,149 0,122 0,186 0,208 0,188
TOTALTON18 -,420** 0,043 -,320* 0,083 ,829** 0,173 0,164 ,319* -0,04 -0,12 ,435** ,399** ,438** 0,062 0,287 ,313* 1 ,903** ,847** 0,092 -,314* 0,129 -0,08
GENERALTON19 -,379* 0,129 -,335* -0,05 ,889** 0,069 0,131 0,18 0,23 -0,01 ,375* ,384* 0,217 -0,05 0,256 ,376* ,903** 1 ,537** 0,048 -0,19 0,154 -0,04
BULKTON20 -,354* -0,08 -0,22 0,22 ,528** 0,254 0,161 ,403** -,358* -0,22 ,391** ,309* ,591** 0,186 0,247 0,149 ,847** ,537** 1 0,12 -,387* 0,063 -0,11
EURPERSON21 -0,03 0,185 0,283 0,273 0,062 -0,14 -0,06 ,384* -0,05 0,119 -0,05 0,002 ,403** -0,04 -0,04 0,122 0,092 0,048 0,12 1 -0,09 -0,1 -0,1
EURTON22 -0,07 0,256 0,149 -0,23 -0,21 -,342* -0,03 -0,3 0,258 ,533** -0,23 -0,28 -,468** -0,21 -0,06 0,186 -,314* -0,19 -,387* -0,09 1 -0,01 0,084
DEABCC23 -0,12 -0,05 -0,13 0,093 0,18 -0,06 0,076 -0,05 -0,05 0,018 0,039 0,156 -0,01 0,184 0,13 0,208 0,129 0,154 0,063 -0,1 -0,01 1 ,791**
DEACCR24 -0,13 -0,04 0,006 -0,03 -0,04 -0,19 0,232 -0,06 -0,1 0,023 -0,12 0,031 -0,02 0,142 0,25 0,188 -0,08 -0,04 -0,11 -0,1 0,084 ,791** 1
*. Correlation is significant at the 0.05 level (2-tailed).
**. Correlation is significant at the 0.01 level (2-tailed).
Factorial Data Analysis

In a factorial analysis of all data, we founded that the ports can by classified in the follow
groups:Group 1 – Port with BULKTON20 <10 Mio tons and GENERALTON19 <5 Mio
tons;Group 2 – Port with BULKTON20 >10 Mio tons and GENERALTON19 <5 Mio
tons;Group 3 – Port with BULKTON20 >10 Mio tons and GENERALTON19 >5 Mio
tons;Group 4 – Port with BULKTON20 <10 Mio tons and GENERALTON19 >5 Mio tons;
To simplify, we call:Group 1 – SP – Small Ports; Group 2 – BP – Bulk Ports; Group 3 –
BMP – Big Multifuntional Ports; Group 4 – GCP - General cargo Ports;

Figure 4

Since some of the variables did not pass the normality test, we used the application of a
nonparametric test to compare the behavior of the averages of all the dependent and
independent variables for each of the four groups of ports defined, the test-Krustkal Wallis.
Table 4
Mean Rank Krustkal-Wallis Test

GENERALTON19
REGULARLSHIP

EURPERSON21
TOTALTON18
PORTPRIV16
MAXDRAFT9
CRAINSKM7

BIGSHIPO15

BULKTON20
SHIPSIZE14
DROTERD2

TERMSIZE8

TXHORIZ11

DEACCR24
GDPCAP17

EURTON22

DEABCC23
SEAPORT4

TXCONT12
TXUNIT10
DMEDIT3

QUAYL6
DCITY5
GRUPES

S13
PEC25 N
1,00 19 27,29 21,95 22,34 22,16 12,97 21,21 16,00 16,21 25,18 19,92 16,34 19,63 14,53 20,32 17,92 19,58 10,00 12,68 12,08 19,03 26,76 21,29 21,13
2,00 7 21,57 11,14 24,93 22,57 25,21 21,71 24,14 33,43 5,57 17,07 27,00 29,14 36,43 23,93 22,29 20,71 29,29 15,71 32,14 24,36 13,00 22,71 23,93
3,00 11 14,18 27,00 18,23 21,73 33,55 25,68 27,73 27,23 20,91 23,45 26,55 21,95 29,55 25,23 29,55 24,14 37,00 35,91 36,00 24,73 13,41 22,36 20,68
4,00 6 20,08 25,67 24,42 21,33 25,67 18,08 28,00 17,42 33,08 31,67 25,75 21,25 15,00 19,17 20,75 27,25 24,00 33,33 15,92 23,67 33,17 22,75 24,92
Total 43
Chi-Square 7,786 7,492 2,666 ,041 20,091 1,609 8,201 12,575 17,964 5,420 7,032 2,964 21,811 1,976 8,160 2,267 35,560 30,598 31,514 1,938 16,245 ,117 ,703
df 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
Asymp. Sig. ,051 ,058 ,446 ,998 ,000 ,657 ,042 ,006 ,000 ,144 ,071 ,397 ,000 ,577 ,043 ,519 ,000 ,000 ,000 ,585 ,001 ,990 ,873

The test result does not reject the hypothesis H0, and at least one group has an average
value different from others, in the variables BULKTON20, GENERALTON19,
TOTALTON18 and TXUNIT10 also for following variables: DROTERD2, DMEDIT3,
QUAYL6, TERMSIZE8, MAXDRAFT9, TXCONT12, SHIPSIZE14, PORTPRIV16 and
EURTON22.
For these variables we can see that the averages are lower the rest for the following groups:
Group 1 – SP – Small Ports – in QUAYL6, TERMSIZE8, MAXDRAFT9, TXCONT12 e
SHIPSIZE14; Group 2 – BP – Bulk Ports – in DMEDIT3, TXUNIT10 e EURTON22;
Group 3 – BMP – Big Multifuntional Ports – in DROTERD2, e EURTON22, bem como
valores elevados em POTRPRIV16; Group 4 – GCP - General cargo Ports – in
MAXDRAFT9 e SHIPSIZE14;
that is, seems to have obtained a model to classify a general European ports on their
specialization and size, with the definition of groups of ports that have different
characteristics with each other.
This could imply that certain features of the ports are connected, determine or be
determined by the specialization of the port and its size, which can be reveled with
intersection of variables BULKTON20 and GENERALTON19.

Figure 5
Groups of Ports
Group 2 – BP – Bulk Ports Group 3 – BMP – Big Multifuntional Port
+

Privatized ports near Rotherdam, deap acess


Ports near Mediterran Sea, deap channel, long quays, beg terminals and low
BULKTON20

acess channel, low prices per tonne prices per tonne


Group 1 – SP – Small Ports Group 4 – GCP - General Cargo Ports
public ports with short quay, short Ports with long quay and big terminals, acess
terminals, acess channel for channel for regional ships, High weigth of
regional ships, high prices per general and roro cargo and high prices per
tonne and low containerization rate tonne
-

- GENERALTON19 +
Phases of the Model Equation
The application of the model equation was carried out in phases for each of the dependent
variables by multiple linear regression with all factors of the model embodied in the
variables that were presented.
To obtain a parsimonious model that explains the results of the ports in terms of its various
dimensions of financial performance and operational efficiency, we used the stepwise and
backward, allowing successively eliminate the variables, analyzing even the hypothesis of
normality , homogeneity and independence of errors, the Durbin-Watson values near 2.
The values Variance Inflation Factor (VIF) for each independet variable of the models,
there are no serious problems of multicollinearity between the factors of the regressions.
To test the hypothesis of a relationship in non-linear format between dependent variables
and some of the explanatory variables were also carried out tests on the logarithmic
variables, with very significant results for the dependent variables BULKTON20 ,
DEACCR24, DEABCC23.

Linear Regression Model for TOTALTON18


In the case of linear regression model for TOTALTON18, the 11th multiple linear
regression, with a Durbin-Watson value of 1.849, helped to identify the variables
DROTERD2, DMEDIT3, QUAYL6, CRAINSKM7, TXUNIT10 and SHIPSIZE14, as
significant predictors of TOTALTON18.
The total cargo of ports depends on the proximity of Rotterdam and the Mediterranean Sea
from, on size, on intensity of cranes per km of quays, on ships size, as well on
specialization in general cargo .
The resulting model is then adjusted TOTALTON18 ^ = -13,189 * DEROTERD2 -
5,393.56 * DEMEDIT3 + 1,833.123 * QUAYL6 + 2,213,756 * CRAINSKM7 +
22,500,000 * TXUNIT10 + 2,086.998 * SHIPSIZE14.
This model is highly significant, with p value <0.05, and explains a high proportion of the
variability of TOTALTON18, with an adjusted R2 = 0.874.
Table 5
Coefficientsa,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
11 DROTERD2 -13189 3787,822 -0,411 -3,482 0,001 0,246 4,07
DMEDIT3 -5393,56 3037,364 -0,188 -1,776 0,084 0,305 3,284
QUAYL6 1833,123 217,199 0,672 8,44 0 0,539 1,855
CRAINSKM7 2213756 867797,9 0,245 2,551 0,015 0,372 2,69
TXUNIT10 2,25E+07 9811366 0,282 2,29 0,028 0,226 4,423
SHIPSIZE14 2086,998 510,455 0,379 4,089 0 0,398 2,511

Linear Regression Model for GENERALTON19


In the case of linear regression model for GENERALTON19 the 8th linear regression, with
a Durbin-Watson value of 1.713, helped to identify the variables DROTERD2, DMEDIT3,
QUAYL6, CRAINSKM7, TXUNIT10 and REGULARSHIPS13, as significant predictors
of GENERALTON19.
The port general cargo, depends on the proximity of Rotterdam or the Mediterranean Sea
on size, on intensity of cranes per km of quays, on regular lines, as well on specialization in
general cargo.
The resulting model is then adjusted GENERALTON19 ^ = = -8,683.23 * DEROTERD2
-3,233.52 * DEMEDIT3 + 1,278.675 * QUAYL6 + 830,430.8 * CRAINSKM7 +
16,900,000 * TXUNIT10 + 772,000,000 * REGULARSHIPS13.
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of GENERALTON19, with an adjusted R2 = 0.864.
Table 6
Coefficients a,b
Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
8 DROTERD2 -8683,23 2011,078 -0,47 -4,318 0 0,268 3,738
DMEDIT3 -3233,52 1695,163 -0,196 -1,907 0,064 0,3 3,332
QUAYL6 1278,675 121,361 0,815 10,536 0 0,53 1,887
CRAINSKM7 830430,8 491433,4 0,159 1,69 0,099 0,356 2,811
TXUNIT10 1,69E+07 5067345 0,367 3,329 0,002 0,26 3,844
REGULARLSHIPS13 7,72E+08 4,57E+08 0,162 1,687 0,1 0,342 2,924

Linear Regression Model for BULKTON20


In the case of linear regression model for BULKTON20 the 13th multiple linear regression,
with a Durbin-Watson value of 1.654, helped to identify the variables DROTERD2,
QUAYL6, CRAINSKM7, SHIPSIZE14, as significant predictors of BULKTON20.
The bulk cargo of ports, depends on the proximity of Rotherdam, port size, intensity of
cranes per km of quays and vessels size.
The resulting model is then adjusted BULKTON20 ^ = -4,192.27 * DEROTERD2 +
530.255 * QUAYL6 + 1,137,184 * CRAINSKM7 + 1,650.35 * SHIPSIZE14.
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of BULKTON20, with an adjusted R2 = 0.697.
Table 7
Coefficientsa,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
13 DROTERD2 -4192,27 2020,259 -0,246 -2,075 0,045 0,502 1,992
QUAYL6 530,255 144,128 0,366 3,679 0,001 0,711 1,406
CRAINSKM7 1137184 596950 0,237 1,905 0,064 0,457 2,191
SHIPSIZE14 1650,35 360,176 0,564 4,582 0 0,465 2,151

In the case of linear regression model with logarithmic variables BULKTON20 for the 14th
regression, with a Durbin-Watson value of 2.115, helped to identify the variables QUAYL6
and TXUNIT10, as significant predictor of BULKTON20.
That is, the ports bulk cargo depends on port size and its specialization in bulk.
The resulting model is then adjusted BULKTON20 ^ = QUAYL6 ^ 1.293 * TXUNIT10 ^
(-2.711)
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of BULKTON20, with an adjusted R2 = 0.885.
Table 8
Coefficientsa,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
QUAYL6 1,293 0,144 0,723 8,958 0 0,41 2,438
TXUNIT10 -2,711 0,822 -0,266 -3,297 0,002 0,41 2,438

Linear Regression Model for EURTON22


In the case of linear regression model for EURTON22 the 13th multiple linear regression,
with a Durbin-Watson value of 2.028, helped to identify the variables TXHORIZ11 and
GPDCAP17, as significant predictors of EURTON22.
That is, the Port Authority revenue per ton depends on the specialization in Roro and region
GPD.
The resulting model is then adjusted EURTON22 ^ = 3.538 * TXHORIZ11 + 0.023 *
GPDCAP17.
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of EURTON22, with an adjusted R2 = 0.711.
However, the model for the EURTON22 included as an explanatory variable the
GPDCAP17, which is a control variable, that is influencing the other independent variables
in the model and it was verified the effect of the elimination of this variable control .
Thus, in the case of linear regression model for EURTON22 without the control variable
GPDCAP17, the 12th multiple linear regression, with a Durbin-Watson value of 2.04,
identified the variables DMEDIT3, MAXDRAFT9 and TXHORIZ11, as significant
predictors of EURTON22.
The revenue per ton of the port Port Authority depends on the degree of specialization in
Roro cargo and on the availability of larger depth in the maritime access to the port and on
the distance to the Mediterranean Sea.
The resulting model is then adjusted EURTON22 ^ = 0.001 * DMEDIT3 + 0.184 *
MAXDRAFT9+ 3.035 * TXHORIZ11
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of EURTON22, with an adjusted R2 = 0.737.
Table 9
Coefficients a,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
DMEDIT3 0,001 0 0,244 1,933 0,06 0,383 2,611
MAXDRAFT9 0,184 0,05 0,73 3,673 0,001 0,154 6,476
TXHORIZ11 3,035 1,07 0,303 2,836 0,007 0,534 1,872

Linear Regression Model for EURPERSON21


In the case of linear regression model for EURPERSON21 the 13th multiple linear
regression, with a Durbin-Watson value of 2.293, helped to identify the variables
DMEDIT3, CRAINSKM7, MAXDRAFT9 and SHIPSIZE14, as significant predictors of
EURPERSON21.
The revenue per employee of the port Authority depends on the distance from the
Mediterranean Sea, intensity of cranes per km of quays, depth in the maritime access to the
port and vessels size.
The resulting model is then adjusted EURPERSON21 ^ = + 0.065 * DMEDIT3 - 15.642 *
CRAINSKM7 + 0.012 * MAXDRAFT9 + 6.903 SHIPSIZE14
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of EURPERSON21, with an adjusted R2 = 0.677.
Table 10
Coefficientsa,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
13 DMEDIT3 0,065 0,022 0,396 2,914 0,006 0,407 2,46
CRAINSKM7 -15,462 7,071 -0,299 -2,187 0,035 0,401 2,493
SHIPSIZE14 0,012 0,005 0,369 2,187 0,035 0,265 3,779
MAXDRAFT9 6,903 3,632 0,41 1,901 0,065 0,161 6,193

Linear Regression Model for DEACCR24


In the case of linear regression model for DEACCR24 the 13th multiple linear regression,
with a Durbin-Watson value of 2.075, has identified the variable with the value
GPDCAP17 B, , as significant predictor of DEACCR24.
That is, a higher relative efficiency DEA CCR of ports, depends on a higher income of the
hinterland.
The resulting model is then adjusted DEACCR24 ^ = + 0.331 * GPDCAP17
This model is significant and explains a high proportion of the variability of DEACCR24,
with an adjusted R2 = 0.487.
However, the model for the DEACCR24 included as an explanatory variable the
GPDCAP17, which is a control variable, that is influencing the other independent variables
in the model and it was verified the effect of the elimination of this variable control .
As a result, the 12th multiple linear regression, with a Durbin-Watson value of 2.17,
identified the variables PORTPRIV16 and SEAPORT4, as significant predictors of
DEACCR24 .
That is, relative efficiency DEA CCR ports depends on the degree of privatization of the
port and the proximity to the sea.
The resulting model is then adjusted DEACCR24 ^ = + 26.357 * PORTPRIV16 +17.882 *
SEAPORT4
This model is significant and explains a high proportion of the variability of DEACCR24,
with an adjusted R2 = 0.467.
Table 11
Coefficients a,b
Standardiz
ed
Unstandardized Coefficient
Coefficients s Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
SEAPORT4 17,882 6,885 0,346 2,597 0,013 0,698 1,434
PORTPRIV16 26,357 7,825 0,449 3,369 0,002 0,698 1,434
In the case of linear regression model with logarithmic variables DEACCR24 for the 14th
regression, with a Durbin-Watson value of 2.22, identified the variable with the value
TERMSIZE8 B, , as significant predictors of DEACCR24.
The efficiency of ports depends on the average size of the port terminals.However, recalls
that many of the authors claim that the indicator DEACCR24 should not apply to ports,
because ports have increasing returns to scale.
The resulting model is then adjusted DEACCR24 = TERMSIZE8 ^ (0.204)
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of DEACCR24, with an adjusted R2 = 0.887.
Table 12
Coefficients a,b
Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
TERMSIZE8 0,204 0,011 0,943 18,375 0 1 1

Linear Regression Model for DEABCC23


In the case of linear regression model for DEABCC23, the 15th multiple linear regression,
with a Durbin-Watson value of 2.039, helped to identify the variables BIGSHIPO15 and
GPDCAP17, as significant predictors of DEABCC23.
That is, a higher relative efficiency DEA BCC port depends on its container shipping lines
global operators and on GPD of the region.
The resulting model is then adjusted DEABCC23 ^ = + 36.965 * BIGSHIPO15 + 0.471 *
GPDCAP17.
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of DEABCC23, with an adjusted R2 = 0.667.
However, the model for the DEABCC23 included as an explanatory variable the
GPDCAP17, which is a control variable, that is influencing the other independent variables
in the model and it was verified the effect of the elimination of this variable control .
Thus, in the case of linear regression model for DEABCC23 without the control variable
GPDCAP17, the 10th multiple linear regression, with a Durbin-Watson value of 2.203,
helped to identify the variables PORTPRIV16, BIGSHIPO15, TXHORIZ11,
REGULARLSHIPS13, as significant predictors of DEABCC23.
That is, relative efficiency DEA BCC port, depends on the degree of privatization of port,
on regular lines and global container shipping lines and on specialization in Roro cargo.
The resulting model is then adjusted DEABCC23 ^ = +22.215 * PORTPRIV16 + 50.374 *
BIGSHIPO15 +36.671 * TXHORIZ11 +4,205.981 * REGULARLSHIPS13
This model is significant and explains a high proportion of the variability of DEABCC23,
with an adjusted R2 = 0.622.
Table 13
Coefficientsa,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
TXHORIZ11 36,671 17,593 0,218 2,084 0,044 0,806 1,241
REGULARLSHIPS13 4205,981 1408,104 0,354 2,987 0,005 0,625 1,601
BIGSHIPO15 50,374 19,092 0,269 2,638 0,012 0,847 1,181
PORTPRIV16 22,215 10,097 0,271 2,2 0,034 0,579 1,727
In the case of linear regression model with logarithmic variables DEABCC23 for the 14th
regression, with a Durbin-Watson value of 2.291, has identified the variable with the value
QUAYL6 B, , as significant predictor of DEABCC23. That is, the more efficient ports
simply depends on port size.
The resulting model is then adjusted DEABCC23 ^ = QUAYL6 ^ (0.41)
This model is highly significant with a p value <0.05, and explains a high proportion of the
variability of DEABCC23, with an adjusted R2 = 0.917.
Table 14
Coefficients a,b

Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
QUAYL6 0,41 0,019 0,959 21,828 0 1 1

Figure 6 – Relations founded


Location
Size
Infrastructure Operational Performance
Specialization
Maritime Services

Location
Infrastructure
Specialization Financial Performance
Maritime Services

Location
Size
Infrastructure Port Eficiency
Specialization
Maritime Services
Logistical Integration
Governance

6. DISCUSSION
It was found that the greater proximity to Rotherdam, the better the performance of port in
terms of operations as regards the total port traffic, general cargo and bulk.
It was also found that the greater proximity to the axis of the Mediterranean, the better the
performance in terms of total port traffic and general cargo and the better is the
performance of the port in financial terms in relation to revenues per ton of cargo and
revenue per employee of the Port Authority.
It was found that the proximity of the sea contributes to a better performance in terms of
efficiency.
There was no relationship between the distance from the harbor to the city and its
performance.
The larger is the size of the port, the better is the performance in terms of port operations as
regards the total port traffic, the general cargo and bulk and the better is the performance of
the port in terms of port efficiency , and even it is the only explanatory factor for the
efficiency indicator on DEABCC23 in the logarithmic model.
The greater the number of cranes per km of quay at the port, the better is the performance
of port operations as regards the total port traffic, the general cargo and bulk, but the worse
is the performance in financial terms of revenue per employee of the Port Authority.
It also was found that the larger depth of access to the port, the better is the performance of
the port in financial terms in respect of revenue per employee of the Port Authority.
The larger is the average size of the port terminals, the better is the performance of the port
in terms of efficiency DEACCR24.
It was found that the higher is the rate of unitization in the port, the better is the
performance in terms of port operations as regards the total port traffic, and general cargo.
The higher is the horizontalization rate at the port, the better is the performance of the port
in financial terms in respect of earnings per ton and the better is the performance in terms
of efficiency with regard to the indicator DEBCC23.
Although the rate of containerization of the port have a significant correlation with the
performance of the port operation, was not considered as an explanatory variable in the
models of port performance, considering that its effect is integrated with other most
important explanatory variables, such as unitization rate, liner and integration with major
global shipping lines.
The larger the average size of the vessels calling at the port, the better the performance in
terms of port operations as regards the total port traffic and bulk.
It was found that the greater the weight of the liner ships to port, the better the performance
in terms of port operations as regards the movement of general cargo, and the better its
performance in terms of port efficiency with indicator DEABCC23.
It was found that the greater the weight of the major global players lines in the regular lines
at the port, the better its performance in terms of port efficiency indicator with
DEABCC23.
It was found that the greater the weight of cargo movement through terminals operated by
the private, the better is the performance in terms of port efficiency indicators for
DEACCR24 and DEABCC23.
The higher the GPD per capita in the region where the port is situated, the better is the
performance in financial terms in relation to revenue per ton and in terms of port efficiency
indicators for DEACCR24 and DEABCC23. However, this effect is indirect, since it is
found that is reflected in the characteristics of the port that have a direct effect on their
performance.

7. CONCLUSION
Implications for Management
This study brings its findings a series of new implications for the management of ports,
including the coordination role of the port authorities, of governments and port terminals
operators.
Firstly it is important to the decision maker to accurately define what exactly is the main
objective of the port or terminal, if if it is to have more total cargo with all the
consequences this has for the region's economy, whether it is to have greater efficiency and
thus have lower costs, making the port an element of support to the competitiveness of
companies in the region or whether it simply wants to increase revenue per tonne or
employee of Port Authority and of all service providers that usually have commissions.
The location of the port or terminal is a basic element to it performance, as found in the
model. But in most cases it is a fact that already exists and is not easy to change, including
the distance to Rotterdam and the Mediterranean Sea. In this case, only the rulers who
decide to create new ports can take these variables into account.
The size of the port is an essential aspect of its operational performance, which is obvious
to any manager. Without a quay there is no cargo and usually without cargo will not be
created a new quay. Seems to be preferable to invest in a large port, which in many
small.The problem is not the understanding of this relationship, but the "timing" in which
they must build a new quay, so that there it will not pass not too many years without
movement to amortize the investment.
The simple increase of the size of the port is not a guarantee of increased port traffic, as
evidenced by several cases. But it appears only if the quay exists. If the relationship
between the size of the port and its cargo is elementary, is more important it relationship
with efficiency due to economies of scale, and its relation to the revenue per tonne and
employee, as a major port should be able to practice lower rates due to economies of scale.
Already the average size of the port terminals is a key variable for the efficiency of the port.
The intensity of use of the quay, with the largest number of cranes, is a decisive factor for
the performance of the ports at the operational level and contributes to lower port costs per
employee, contributing to their competitiveness.
Depth of port access, allowing the port to receive larger vessels, is a key variable for the
performance of ports, including total cargo and bulk, but also for the container market.
However, the large investment required can make highest port charges per ton and per
employee, making the port less competitive, which is certainly offset by the increased
competitiveness of the freight of the larger ships that can reach the port and have a lower
consumption per tonne / km in its travels, giving the advantage to port.
Specialization in general cargo is also an issue that is important for ports that want to have
a higher performance in the total cargo and general cargo.
Attracting more regular lines for the port requires the establishment of management
conditions and appropriate infrastructure, but ensure great performance in terms of the
general cargo, including containers, and requires the port to adjust the degree of relative
efficiency in order to have lower costs.
The study demonstrated the importance of the privatization of port management of the port
terminals, since such a measure contributes significantly to the improvement of port
efficiency and thus its competitiveness, which associated with integration the port in
international networks of major container shipping operators, facilitates the attainment of
performance targets of the port.
Finally, it is surprising conclusions about the importance of specialization rate in roro
cargo, involving performances in terms of higher revenue per tonne, making the port less
competitive, but are advantageous in terms of less need for investment in equipment cranes,
making the port more efficient, with advantages in the speed with which cargo is loaded
and unloaded of the ship.

Research for the Future


Despite the ambition of this study, the result leaves more questions than answers to the
problems of the ports. It is to know the precise extent to which the characteristics of the
ports influence their performance and to measure accurately how the performance of ports
influences the characteristics, since it is certainly a reciprocal relationship over time, which
gradually characteristics of ports and the environmental variables will influence the
performance of ports and this is reflected in contributing to the change of environment of
the port and the characteristics of the port.
Secondly, to verify whether the constructs indicated in this study can be classified as
features of ports, or if some are environmental variables or dependent on other results. It
would be important to define the extent to which factor is determined to each other, how
they relate, and how they contribute directly and indirectly to the performance of the port.
We proceed with a possible model for this, drawn from the correlation coefficients of
Pearson and the results of the study.

Figure 7

Location Infrastructure
Governance

Specialization
Port Perfomance

Region Perfomance Maritime Service


Size
n+1

Finally, several issues may still be investigated and further developed, such as the type of
management of terminal, the port terminals characteristics that determine then performance,
what determines the performance of the port at the level of service to ships and waiting
times.
Also in the present study, may also be asked several questions such as, for example: The
largest ports with more private management have better performance?The ports with larger
terminals and larger depths have better performance?
The relationship between the variables that characterize the port and between them and
their performance could be assessed by a survey of the opinion of major European players
in the port sector, custmers and operators.

Strengths and Limitations of This Analysis


This study's main limitation is the number of responses that was obtained from european
port authorities, who collaborated with difficulty, with partial responses, which were
completed with great effort and persistence.Nevertheless, it was not possible to feed the
database of the sample with all variables initially intended, as some of the issues have not
received a response from some ports, making it impractical to use, despite the possible
importance for the model.
The conclusions of such studies are importante for ports, so it would be important that the
port authorities collaborate more actively in the responses to surveys.
An alternative is ESPO to promote a annual survey to collect a wide range of information
on European ports and make available to all researchers wishing to study the ports.
Nevertheless, given that the vast majority of studies on the ports are restricted to the ports
of certain countries or to very simple and basic quantitative data or to ESPO and published
by Eurostat, one of the strengths of this study was able is to cover a wide range of
quantitative variables in a large population of European ports, which allows a better
understanding of diversity of characteristics that determine the performance of ports.
REFERENCES

Ahn, T., Charnes, A., Cooper, WW (1988), "Some Statistical and DEA Evaluations or
Relative Efficiencies of Public and Private Institutions of Higher Learning," Socio-
Economic Planning Sciences, Vol 22 pp.259 -- 269;
Ashar, A., (1997), "Counting the Moves," Port Development International, 13: PP25-29;
Badin, NT, (1997), "Productivity of supermarkets and their benchmarking, thesis was
approved as a partial requirement for obtaining a Masters in the Post-graduate in
Production Engineering from Universidade Federal de Santa Catarina, 1997;
Banker, RD, Charnes, A. and Cooper, W., (1984), "Models for Estimation of Technical and
Scale Inefficiencies in Data Envelopment Analysis", Mgmt. Sci 30, 1078-1092.Novaes
(1997 and 2001);
Barros, CP (2003), "Incentive Regulation and Efficiency of Portuguese Port Authorities",
Maritime Economics & Logistics, 2003, 5, (55-69);
Barros, CP (2006), "A Benchmark Analysis of Italian Seaports Using Data Envelopment
Analysis", Maritime Economics & Logistics, 2006, 8, (347-365);
Barros, CP and Athanassiou, M., (2004), Efficiency in European Seaports with DEA:
Evidence from Greece and Portugal ", Maritime Economics & Logistics, 2004, 6, (122-
140);
Barros, CP, and Peypoch, N., (2007), "Comparing Productivity Change in Italian and
Portuguese Seaports using the Luenberger Indicator Approach," Maritime Economics &
Logistics, 2007, 9, (138-147);
Bendall, H., Stent, A., (1987), "On Measuring Cargo Handling Productivity", Maritime
Policy and Management, 14: pp.337-343.
Boussofiane, A., Dyson, RG and Thanassoulis, E., (1991), "Applied Data Envelopment
Analysis", European Journal of Operational Re search, vol 52, No. 1, pp.1-15;
Caldeirinha, V., 2007, "Efficiency and Competitiveness of Ports", Publication of the
Seminar, Instituto Superior Técnico, Lisbon;
Chang YT, Lee and Paul TW (2007), "Overview of interport competition: Issues and
methods", Journal of International Trade and logistics, 99, Volume 5, Number 1, June
2007, pp.99 ~ 121;
Chang, YT, and Lee, PT, (2007), "Overview of interport competition: Issues and methods",
Journal of International Logistics and Trade, 99 Volume 5, Number 1, June 2007, pp.99-
121;
Charnes, A., Cooper, W. and. Rhodes, E., (1978), "Measuring the Efficiency of Decision
Making Units", Eur J. Opnl.Res 429-444;
Cheon, S. (2007), "Evaluating Impacts of Institutional Reforms on Port Efficiency Changes
malquimist Productivity Index for World Container Ports," Post Doctoral Research,
University of California, Berkeley;
Chu XH; Fielding GJ, Lamar BW (1992), "Measuring transit performance using data
envelopment analysis", Transportation Research Part A-Policy and Practice, 26 (3): pp223-
30;
Coto-Millán, P., Baños-Pino, J. y Rodríguez-Álvarez, A. (2000): "Economic Efficiency in
Spanish Ports: Some Empirical Evidence", Maritime Policy and Management, 27 (2), 169-
174;
Christopher Dos Santos, RA, and Haddad, EA, (2007), "Relative Efficiency of Brazilian
Ports: An Analysis Regionalized, IPE-USP, FIPE, REAL-UIUC, Brazil;
Culinnane, K., et al. (2004), "An Application of DEA Windows Analysis to Container Port
Production Efficiency," Review of Network Economics Vol.3, Issue 2 - June 2004;
Culinnane, K., et al., (2005), "The Application of Mathematical ProgrammingApproaches
to Estimating Container Port Production Efficiency," Journal of Productivity Analysis, 24,
73-92, 2005;
Cullinane, K., Song, D. (2002), "Port privatization policy and practice", Transport Reviews,
Vol 22 No.1;
Cullinane, KPB & Song, DW., (2002), "Port Privatization: Principles and Practice",
Transport Reviews 2002, 22 (1), 55-75;
Cullinane, KPB, (2002), "The Productivity and Efficiency of Ports and Terminals: Methods
and Applications", In: CT Grammenos, ed. The Handbook of Maritime Economics and
Business. London: Informa Profession, 2002, pp.803-831;
Cullinane, KPB, Ji, P. & Wang, T., (2002), "A Multi-Objective Programming Approach to
the Optimization of China's International Container Transport Network", International
Journal of Transport Economics 2002, XXIX (2), 181-199;
Cullinane, KPB, Song, DW. & Gray, R. (2002), "A Stochastic Frontier Model of the
Efficiency of Major Container Terminals in Asia: Assessing the Influence of
Administrative and Ownership Structures," Transportation Research A 2002, 36 (8), 743-
762;
Cullinane, KPB, Li, KX & Cheng, J. (2003), "The Application of WTO Rules in China and
the Implications for Foreign Direct Investment", Journal of World Investment 2003, 4 (2),
343-361;
De Monie, G. (1987), "Measuring and Evaluating Port Performance and Productivity",
UNCTAD Monographs on Port Management No. 6 on Port Management, Geneva:
UNCTAD;
De-Neufville, R and Tsunokawa, K. (1981), "Productivity and returns to scale of container
port", Maritime Policy and Management, Vol 8, No. 2, pp121-129;
Díaz-Hernández, JL, et al., (2007), "Productivity in Cargo Handling in Spanish Ports
During a Period of Regulatory Reforms," Netw Spat Econ (2008) 8:287-295;
E. Thanassoulis, (1993), "A comparison of regression analysis and data envelopment
analysis as alternative methods for performance assessments", Journal of the Operational
Research Society, Vol 44, No 11, pp.1129-1144;
Estache, A, et al. (2001), Technical Efficiency Gains from Port Reform: The Potential for
Yardstick Competition in Mexico ", The World Bank Institute, Governance, Regulation,
and Finance Division;
Estache, A., et al., (2005), "Infrastructure Performance and Reform in Developing and
Transition Economies: Evidence from a Survey of Productivity Measures", World Bank
Policy Research Working Paper 3514, February 2005;
Estrada, José Luis, (2007), "Mejora de la competitividad de puerto by an average of one
new model a la gestión strategy framework for applying el mando integral", Universidad
Politécnica de Madrid;
Fleming, DK, Baird, AJ (1999), "Comment Some reflections on port competition in the
United States and western Europe", Maritime Policy & Management, Volume 26, Number
4, 1 October 1999, pp. 383-394 (12);
Garcia-Alonso, L., and Martin-Bofarull, M., (2007), "Impact of Port Investment on
Efficiency and Capacity to Attract Traffic in Spain: Bilbao versus Valencia", Maritime
Economics & Logistics, 2007, 9, (254 -267);
Gaur, P. (2005), "Port Planning as a Strategic Tool: A Typology, Institute of Transport and
Maritime Management Antwerp, University of Antwerp (April 2006);
Giokas, DI, (1990), "Bank Branch operating efficiency: a comparative application of DEA
and the loglinear model", Omega International Journal of Management Science
Gonzalez, MM, and Trujillo, L, (2007), "Efficiency Mesurement in the Port Industry: a
Survey of Empirical Evidence", City University, London;
Goss RO, (1990), "Economic policies and seaports: The diversity of port policies,"
Maritime Policy & Management, Vol 17, No. 3. (1990), pp.221-234;
Guthed, Arvid (2005), "Port Hinterland Connections", Division of Transportation and
Logistics, Chalmers University of Technology, Göteborg, Sweden;
Guy, E., Urli, B., (2006), "Port Selection and Multicriteria Analysis: An Application to the
Montreal-New York Alternative," Maritime Economics & Logistics, 2006, 8, (169-186);
Haralambides, H. (2002), Competition, Excess Capacity and Pricing of Port Infrastructure,
International Journal of Matitime Economics, (2002), 4, (323-347);
Haralambides, He, et al., (2003), "Port Financing and Pricing in the European Union:
Theory, Politics and Reality", International Journal of Maritime Economics, 2001, 3, (368-
386);
Herrera, S., and Pang, G., (2006), "Efficiency of Infrastructure: The Case of Container
Ports";
Hui, E., Seabrooke, W., and Wong, G., (2004), "Forecasting Cargo Throughput for the Port
of Hong Kong: Error Correction Model Approach," Journal of Urban Planning and
Development, Vol 130, No. 4, December 1, 2004;
Jara-Diaz, S., et al. "Multioutput analysis of cargo handling firms: An application to a
Spanish port", Transportation (2005) 32: 275-291;
Jose L. Tongzon, (2000), "Determinants of port performance and efficiency,"
Transportation Research Part A: Policy and Practice, Volume 29, Issue 3, May 1995, Pages
245-252;
Kent, P., and Ashar, A., (2001), "Port Competition Regulation: A Tool for Monitoring for
Anti-Competitive Behavior," International Journal of Maritime Economics, 2001, 3, (27-
51);
Kerlinger, FN, (1980), "Analysis of covariance structure: tests of Criteria1 referents
theory", Multivariate Behavioral Research 15: 403-422;
Kim, Moshe & Sachish, Arie (1986), "The Structure of Production, Technical Change and
Productivity in a Port", Journal of Industrial Economics, Blackwell Publishing, vol. 35 (2),
pages 209-23, December;
Langen, Peter De, (2004), "Governance in Seaport Clusters", Maritime Economics &
Logistics, 2004, 6, (141-156);
Laxe, FG, (2005), "The Port Competitiveness Indicator Through the Multicriteria Decision
Method Promethee, A Practical Implementation to the Spanish Port System, Spanish
Ministery of Public Works;
Lee, HS et al., (2005), "Evaluating Port Efficiency in Asia Pacific Region With Recursive
Data Envelopment Analysis", Journal of the Eastern Asia Society for Transportation
Studies, Vol 6, pp. 544-559, 2005;
Lee, Sung (2006), "Performance Evaluation of Asian Port Distriparks Using Factor
Analysis";
Liu, BL, et al., (2005), "Efficiency Analysis of Container Terminals in China: An
Application of DEA Approach", Institute of Transportation Economics, Nankai University,
Tianjin, China;
Liu, Zinan (1995), "The Comparative Performance of Public and Private Enterprises: The
Case of British Ports," by The London School of Economics and Political Science and
University of Bath;
Serrano María Manuela González & Lourdes Trujillo Castellano, (2005), "La medición de
la eficiencia en el sector portuario: review of empirical evidence," Documentos de trabajo
set ULL-ULPGC 2005-06, Facultad de Ciencias Económicas de la ULPGC;
Marlow, P., and Passion, AC (2003), "Measuring lean ports performance", International
Journal of Transport Management 1 (2003) 189-202;
Moroco, John (2003), "Statistical Analysis using SPSS", published Silabo, Lisbon;
Martinez, BE, Diaz, AR, Navarro, IM; Ravelomesa, T., (1999), "A Study of the Efficiency
of Spanish port authorities using Data Envelopment Analysis", International Journal of
Transport Economics, 1999, Vol 26, N . 2, pp. 237-253.
Medda, F., and Carbonaro, G. (2007), "Growth of Container Seaborne Traffic in the
Mediterranean Basin: Outlook and Policy Implications for Port Development", Transport
Reviews, A Transnational Transdisciplinary Journal, 7 June 2008;
Moita, MH (1995), Measuring the Relative Efficiency of municipal schools of Rio Grande -
RS Approach Using DEA, Florianopolis;
Moita, M., (1995), Measuring the Relative Efficiency of municipal schools of Rio Grande
DEA "Universidade Federal de Santa Catarina, Master's thesis;
Morey, RC, Fine, DJ and Loree, SW, (1989), "Comparing the Allocative efficiencies of
hospitals", OMEGA Int J Manage. Sci 18 (1) :71-83, 1989;
Nchez, R., et al., (2003), "Port Efficiency and International Trade: Port Efficiency as a
Determinant of Maritime Transport Costs," Maritime Economics & Logistics (2003) 5,
199-218;
Neter, J., Wasserman, W. & Kutner, MH (1985), "Applied Linear Regression Models",
Irwin, Homewood, Illinois.Maroco, 2007;
Ng, Ada Suk and Lee, Chee Xui, (2006), "Port productivity analysis by using DEA: A case
study in Malaysia", Institute of Transport and Logistics Studies, The Australian Key Center
in Transport Management, The University of Sydney, Australia;
Notteboom Theo, Coeck Chris, Julien van den Broeck, (2000), "Measuring and explaining
the relative efficiency of container terminals by means of Bayesian stochastic frontiers
models" - In: International Journal of Maritime Economics, 2:2, (2000) ECR 83-106;
Notteboom, TE and W. Winkelmans (2001) "Structural changes in logistics: how will port
authorities face the challenge?", Maritime Policy and Management, 28:71-89;
Notteboom, Theo E., and Rodrigue, JP, (2005), "Port Regionalization: Towards a New
Phase in Port Development, Maritime Policy &Management;
Odeck, J., Hjalmarsson, L. (1996), "Performance of trucks-an evaluation using data
envelopment analysis", Transportation Planning and Technology, 20 (1): pp.49-66.
Park, RK, De, P., (2004), "An Alternative Approach to Efficiency Measurement of
Seaports," Maritime Economics & Logistics, 2004, 6, (53-69);
Poitras, G., et all., (2005), "Measuring Port Efficiency: An Application of Data
Envelopment Analysis", Department of Economics and Statistics, National University of
Singapore;
Poitras, G., Tongzon, J. and Li, H., (1996), "Measuring Port Efficiency: An Application of
Data Envelopment Analysis", Department of Economics and Statistics National University
of Singapore;
Rios, LR, et al. (2006), "Measuring the Efficiency of Operations of Container Terminals in
Brasil", FURG;
Robinson, R. (2002): "Ports as elements in value-driven chain systems: the new paradigm,
Maritime Poli cy and Management.29: 241-255;
Robinson, R. (2003): "Port authorities: defining functionality within a value chain-driven
paradigm, Proceedings IAME Conference.Busan: 654-674;
Rodriguez-Alvarez, Ana, Tovar, Beatriz, Trujillo, Lourdes, (2005), "Firm and time varying
technical and Allocative efficiency: An application to port cargo handling firms",
International Journal of Production Economics;
Rodrigue, JP, Comtois, C. , Slack, B., (2009), "The Geography of Transport Systems,
Routledge;
Sachish, A., (1996) "Productivity functions as a managerial tool in Israeli ports", Maritime
Policy and Management 23 4, pp. 341-369;
Sanchez, R., Hoffmann J., Micco, A., Zzolitto, G., Sgut, M. & Wilmsmeier, G., (2003),
"Port Efficiency and International Trade: Port Efficiency as a Determinant of Maritime
Transport Costs", Austral University, Argentina, Maritime Economics & Logistics, (2003),
5, (199-218);
Simme J Veldman and Ewout Bückmann H (2003), "A Model on Container Port
Competition: An Application for the West European Container Hub-Ports", Maritime
Economics & Logistics (2003) 5, 3-22;
Simme, JV, and Ckmann, EH Bu, (2003), "Model on Container Port Competition: The
Application for the West European Container Hub-Ports", Maritime Economics &
Logistics, 2003, 5, (3-22);
So, S. et al. (2007) "Efficiency Analysis and Ranking of Major Container Ports in Northeast
Asia: An Application of Data Envelopment", International Review of Business Research
Papers Vol.3 No.2 June 2007, pp. 486 - 503 486;
Song, DW, and Yeo, KT, (2004), "A Competitive Analysis of Chinese Container Ports
Using the Analytic Hierarchy Process," Maritime Economics & Logistics, 2004, 6, (34-52);
Song, DW. & Cullinane, KPB, (2002), "Port Privatization: A New Paradigm of Port
Policy", In: E. Borgese, A. Chircop and M. McConnell [eds.], Ed.Ocean Yearbook.
Chicago: University of Chicago Press, 2002, pp.398-420;
Tabernacle, JB (1995), "A Study of the Changes in Performance of Quayside Container
Cranes", Maritime Policy and Management, 22: pp.115-124;
Tongzon, J., (2001), "Efficiency Measurement of Australian select an International Port
using Data Envelopment Analysis", Transportation Research Part A 35, Elsevier, 2001, p.
113-128;
Tongzon, J., (2002), "Port Choice Determinants in a Competitive Environment", IAME,
Conference Proceedings, 13 - 15 November 2002, Panama;
Tongzon, Jose, and Heng Wu, (2005), "Port privatization, efficiency and competitiveness:
Some empirical evidence from container ports (terminals)," Transportation Research Part A
39 (2005) 405-424;
Trujillo, L., and Tovar, B., (2007) "The European Port Industry: An Analysis of its
Economic Efficiency," Maritime Economics & Logistics, 2007, 9, (148-171);
Turner, H., et al., (2004), "North American containerport productivity: 1984-1997",
Transportation Research Part E 40 (2004) 339-356;
Ugboma, C. et all., (2006), "An Analytic Hierarchy Process (AHP) Approach to Port
Selection Decisions - Empirical Evidence from Nigerian Ports", Maritime Economics &
Logistics, 2006, 8, (251-266);
Valentine, VF, Gray, R., (2001), "The Measurement of Port Efficiency Using Data
Envelopment Analysis", Proceedings of the 9th World Conference on Transport Research,
Seoul, South Korea, 22-27 July;
Wang, TF and Cullinane, K., (2006), "The Efficiency of European Container Terminals and
Implications for Supply Chain Management", Maritime Economics & Logistics, 2006, 8,
(82-99);
Wiegmans, B., (2003), "Performance Conditions for Container Terminals", Maritime
Economics & Logistics (2004) 6, 276-277;
Winkelmans, W, (2003), "Port Competitiveness and Port Competition: Two of a kind?",
Paper for IAPH conference;
Yanbing, Y., Zhongzhen, Y., (2005), "Evaluation of competition ability and market share
for container port", Proceedings of the Eastern Asia Society for Transportation Studies, Vol
5, pp. 2483 - 2493, 2005;
Yeo, GT, and Song, DW, (2006), "An application of the hierarchical fuzzy process to
container port competition: Policy and strategic implications", Transportation (2006) 33:
409-422.