FDI & Ports - Paper Answers

Remarks xfreightstata.dtaRegionalDataEurope.dtaPlan FDIANDPORTSPROJECT Research_Project FDI.dta

Master′s thesis: -STATA programme is needed for the econometric models. -Data are available but modification is needed. -Panel database including regional variables (NUTS-2) level for the period 2003-2010, for Europe. -A combination of the 2 uploaded pdf docs is preferred. -A Stata Dotfile with the whole procedure is also needed

Some thoughts and remarks

Firstly, the abstract part is huge and I think it is more an introduction. Also, the introduction part it could be incorporated in the literature review part.

Secondly, as for the hypotheses there are some parts of theory that seems to me misplaced.

For example, in hypothesis 2(The size of a port affects the number of inward FDI positively) you are referring to GDP a lot. Although, GDP is a very important variable for examining FDI attractiveness in a country I am not sure if GDP should be analyzed in this part which has to do about the size of a port.

Also, in hypothesis 3(The presence of a port in a region and the presence of good infrastructure affect the number of inward FDI positively) I think it should have more references about the relation between the infrastructure of a region and a port and how it affects FDI.

Likewise, in Hypothesis 4(Different types of FDI and if there is a difference between port and no port regions) even though you have correctly analyze the different types of FDI I think it would be better if you could mention market seeking and resource seeking FDI. (Except if you aim to analyze it somewhere else in the assignment) Also, I don’t understand the Irish model and the Singapore story and how they contribute to this hypothesis. Furthermore, the citations are missing in this part.

Typically, the project covers all aspects of doing empirical research with real world

data including finding additional literature for your theoretical framework,

constructing a dataset, conducting statistical analyses, and eventually writing the

research report. Note that you start with one main hypothesis, but that you can

come up with several sub-hypotheses.

HYPOTHESES:

Hypothesis 1 : The presence of a port in a region affects positively the number of

inward FDI.

Hypothesis 2 : The size of a port affects positively the number of inward FDI.

Hypothesis 3 : The presence of a port in a region and the presence of good

infrastructure affects positively the number of inward FDI.

Hypothesis 4 : Different types of FDI and if there is a difference between port and

non-port regions. (for example manufacturing fdi versus services fdi, or Resource

seeking vs market seeking etc, )

-All analyses you have to conduct should be carried out in Stata.

-You can either estimate a fixed effects/random effects/Pooled OLS model or count

data model (Poisson, negative binomial, zero-inflated).

This study investigates the effect of the
presence of a port on the attractiveness of
inward FDI in the region. Port regions will
be defined due its geographical location

and categorized by size. Afterwards, using a
negative binomial regression model,

several related hypotheses will be tested

The results show a positive significant

relationship between a port region and the
inward FDI.

The Effect of
a Port on the
inward FDI in
a region

Table of Content

1. INTRODUCTION

………………………………………………………………………………………………………….3

2. THEORETICAL BACKGROUND

…………………………………………………………………………………..5
2.1 Port regions and FDI ………………………………………………………………………………………………………….

2.2 Positive effects of port on the city …………………………………………………………………………………………

2.3 Business sectors and proximity to port ………………………………………………………………………………….

3. THE HYPOTHESES

……………………………………………………………………………………………………

4. DATA & METHODOLOGY

…………………………………………………………………………………………

4.1 Data ……………………………………………………………………………………………………………………………….. 14
4.2 Ports and urban attractiveness …………………………………………………………………………………………..

4.3 Sectoral Analysis ……………………………………………………………………………………………………………… 16
4.4 Effect of Port size ……………………………………………………………………………………………………………..

4.5 Control Variables ……………………………………………………………………………………………………………..

4.6 Estimation Strategy …………………………………………………………………………………………………………..

5. RESULTS

…………………………………………………………………………………………………………………..

6. DISCUSSION & CONCLUSION

……………………………………………………………………………………

REFERENCES

……………………………………………………………………………………………………………….

APPENDIX

…………………………………………………………………………………………………………………….

1. INTRODUCTION

In the ceaseless game of dominance, multinational companies (MNCs) constantly pursue

competitive gains and advantages in order to secure sustainability. In a fast changing and

demanding global community, firms aim to expand beyond the boarder of their own country,

investing in different geographical areas around the globe, a strategy commonly refer to as

foreign direct investment (FDI). As simple as this may sound, the process of identifying

opportunities abroad is complex and usually involve businesses to analyze different factors

before committing to a specific location (Dunning J. H., 2001) (Porter, 2000). For businesses

interested in investing in European regions, ports could be of pivotal importance. Ports serve as

vital economic gateways as 74% of goods imported in or exported from Europe go by sea.1

European ports are not considered a homogenous set of ports (Notteboom, 2010) not only

because of the varied types of commodity handled but also for reasons related to connectivity

with various hinterlands and different location qualities. With many different ports within a

relatively small continent (Europe), competition to draw FDI becomes fierce among regions.

Another interesting aspect of the European region has to do with the formation of the European

Union. One of main purposes for forming the Union was to facilitate investments and potentially

increase wealth across the Union. The free movement of capital, goods and persons within the

Union as well as no trade barriers and tariffs within the E.U in theory make FDI very appealing

to those firms looking for opportunities abroad. In their study, Bevan & Estrin analyze the

determinants of FDI in European economies and found that “FDI is positively related to both

1 https://ec.europa.eu/transport/modes/maritime/ports/ports_en

https://ec.europa.eu/transport/modes/maritime/ports/ports_en

source and host country GDP and related inversely to the distance between countries and to unit

labor costs” (Bevan & Estrin, 2004). An example of this could be a German firm looking for

opportunities to reduce labor costs in the Eastern part of Europe2. From previous research, we

know that many parts of Europe draw substantial amount of FDI in both services and

manufacturing sectors, the latter sector being explored more often especially in Eastern regions

(Disdier & Mayer, 2004). Furthermore, firms interested in investing in the European regions in

general do not seem to discriminate on the determinants of FDI before committing to a specific

location. This means that market size and agglomeration effect are considered equally important

for all EU regions (Disdier & Mayer, 2004). Disdier and Mayer also mentions that the

competition for FDI is not across regions (West vs East for example) but within the regions itself

(East vs East countries and West vs West countries). An explanation for this might be the

economic characteristics of the region drawing similar FDI interest within these areas.

The economic attractiveness as well as the geographical composition of European regions along

with the presence of many ports within this specific area makes it very interesting for economist

to research. Our research question is therefore if the presence of a port plays an active role in

urban competiveness and the attractiveness of inward FDI? The method used to answer this

question is the so-called “negative binomial model”; a model which is very suiting in order to

control for overdispersion in our data.

This paper continues as follows: In Section 2 we provide the reader with the necessary theory to

understand the dynamics of ports and FDI within the European region as well as our motives for

formalizing the different hypotheses within this paper. Section 3 introduces the composition of

2Data suggests that income in Germany is higher than in for example Bulgaria creating possibly certain advantages.

See:http://databank.worldbank.org/data/Views/Reports/ReportWidgetCustom.aspx?Report_Name=CountryProfile&

Id=b450fd57&tbar=y&dd=y&inf=n&zm=n&country=DEU and

http://databank.worldbank.org/data/Views/Reports/ReportWidgetCustom.aspx?Report_Name=CountryProfile&Id=b

450fd57&tbar=y&dd=y&inf=n&zm=n&country=BGR

http://databank.worldbank.org/data/Views/Reports/ReportWidgetCustom.aspx?Report_Name=CountryProfile&Id=b450fd57&tbar=y&dd=y&inf=n&zm=n&country=DEU

http://databank.worldbank.org/data/Views/Reports/ReportWidgetCustom.aspx?Report_Name=CountryProfile&Id=b450fd57&tbar=y&dd=y&inf=n&zm=n&country=BGR

our dataset, reasons for our choice of regression models and estimation strategy. Section 4 is

reserved for the empirical analysis and results of the main and sub hypotheses. Finally, in

Section 5 we discuss some of the limitations of this paper and conclude accordingly.

2. THEORETICAL BACKGROUND

2.1 Port regions and FDI

There are many definitions of what a port is but for simplicity we will use the definition of

Stopford (2009:81) a port is “a geographical area where ships are brought alongside land to load

and discharge cargo – usually a sheltered deep-water area such as a bay or river mouth”

(Stopford, 2009) (Nijdam & van der Horst, 2018). From this we can already get a “feel” that

ports function as node in transport chains and are important for economic activities involving

cargo and ship handling (Nijdam & van der Horst, 2018). Ports generally are an economic

catalyst for surrounding cities in the region, facilitating the integration of markets and the

agglomeration of services that generate economic benefits and socioeconomic welfare (Song &

van Geenhuizen, 2014) (Zhao, Xu, Wall, & Stavropoulos, 2017). The following paths in spatial

distribution show changes in the economic relationship between ports and port cities. Port cities

usually benefit from the port’s economic activities, by needing to be near the port. An example

of this is the lower transaction costs provided by ports to port related business activities.

Urban spaces near ports also provide ports with advantages that cannot be easily accessed

outside of urban agglomerations, such as labor pools and infrastructure, in this case the city

provides the port the human capital to efficiently run its labor and provide for roads to reach the

hinterland and vice versa (Hall & Jacobs, 2012). Port-related industries are attracted by such

environments, which allows ports and port cities have a relationship with one another and

economically benefit off each other (Zhao, Xu, Wall, & Stavropoulos, 2017). Because the

maritime transport network covers about 90% of world trade (Ducruet, Rozenblat, & Zaidi,

2010), it also has an effect on the global economy. Port and city networks become related to each

other during this process. Jacobs et al. (2010) stated that port-related advanced producer services

activities predominantly follow the overall global city trends, where some port cities have an

advantage over others because they act as hubs in global (commodity) flows on top of acting as

centers of advanced services related to shipping and port activities. But why is one urban port

area more competitive and/or attractive than another? First we need to understand the concept of

urban competitiveness. Kostiainen (2002) states that ‘the ability to attract flows of information,

technology, capital, culture, people, and organization is the key concept of urban

competitiveness (Kostiainen, 2002)’. Urban competitiveness can be measured by the capacity of

cities to attract investment and to promote development (Sáez & Periáñez, 2015). It is important

to note that there are two different types of FDI’s related to ports; inward FDI and outward FDI,

this is in our interest to analyze because inward and outward investment should be separated

from each other due to the different requirements of the involved parties making the investments

(Dooms, Lugt, & Langen, 2013) (Kolstad & Wiig, 2012).

Inward investments give the opportunity to attract international private multinationals, where the

host country can grow investments and provide for local economic growth. Outward investments

aim for outward activities (abroad), and look for exploiting new business opportunities and

relationships (Dooms, Lugt, & Langen, 2013). The port’s (authority) outward

internationalization strategies are to sell the port worldwide (customer seeking principle), create

value for their domestic customers and/or maximize profits through involvement in exploiting

market opportunity abroad (Dooms, Lugt, & Langen, 2013). However, because this study

focuses on the capacity to attract foreign investments in Europe, we solely analyze inward FDI

made within the mentioned region. The ability to attract (inward) FDI is a good indication of a

city’s competitive success (Lovering, 2003). Inward investments allow flow of goods, capital,

resources, information and or services from external world to domestic market (host

country/region/city/port) (Karlsen, Silseth, Benito, & Welch, 2003). The port inward FDI

operations aim to attract more foreign direct investment, attract international private companies,

increase investments and traffic volumes and this all to aid the local economic growth.

2.2 Positive effects of port on the city

With the presence of a port in/near a city, the port has an influence on the city to focus on export

related industries. The transport costs are affected by the connectivity between inland countries

(region) and available ports. For example, in comparison with other inland countries, the Czech

Republic, Switzerland and Austria that are surrounded by ports in the European port system, and

this advantage affords these countries more negotiation power to decrease transport costs in

business opportunities (Merk & Hesse, The Competitiveness of Global Port-Cities: The Case of

Hamburg, Germany , 2012). Secondly, another benefit of ports on urban regions is the creation

of an additional added-value. For example, Rotterdam generated 12.8 billion U.S dollars of

added-value in 2007, which accounted for ten percent of its regional GDP (Merk, 2014) This

value comes from four sources (Ferrari, Parola, & Gattorna, Measuring the quality of port

hinterland accessibility: The Ligurian case, 2011):

 The increase of employment opportunities and income by the construction and operation

of port infrastructures

 The increase of employment opportunities and income by port related industries

 A stimulated domestic demand by the increase of income

 Foreign investment that is attracted by a port’s welfare

Merk (2014) also states that as the scale of a port becomes larger, the added value also grows e.g

up to ten percent in employment in the ports of North West Europe. Thirdly, the growth of (port)

employment: a study of the European port region shows that 100 million units of cargo

throughput will create 0.0003% of regional employment opportunities (Ferrari, O., Bottaso, &

Tei, 2012). To add to this, a port city is becoming an innovation center for port related industries

(2014). Fourth, the spillovers (technology, salary) of economic benefits will spread to other cities

in the vicinity of the port (region). A port can do this by increasing its port competitiveness in

terms of connectivity and port efficiency; the purpose of this is to improve its locational status in

global port networks and thereby increase its urban competitiveness, thus building a bridge

between ports and municipalities (Zhao, Xu, Wall, & Stavropoulos, 2017). In the case study of

Rotterdam, the positive spillovers of the Rotterdam port have even spread to nearby countries

such as German industries (Merk, 2014). Based on this research of Merk (2014) there are three

aspects of port competitiveness. The first one is maritime connectivity, the aspect to measure the

level of accessibility of a certain port. The second aspect is port efficiency, throughput of

container and bulk goods, as well as ship calling and other indicators of port activeness. The

third aspect is hinterland connectivity, how well a port is connected and reachable for the region

around the port.

In order to satisfy these demands of multinational corporations, the destination of foreign

investment should possess as many as attracting factors related to locational advantages. To

emphasize, because we focus on the inward investment, we should focus on the determinants of

investing in the host country.

2.3 Business sectors and proximity to port

In the previous section we discussed the attracting factors of port regions and cities but the

question remains; which business sectors are attracted the most by the presence of a port in an

area? Some business sectors (manufacturing, logistics, transport for instance) are attracted by

being near a port. Being close to a port provide these firms with advantages concerning

transaction and transportation cost (Jacobs, Ducruet, & De Langen, Integrating world cities into

production networks: The case of port cities, 2010). In recent times, ports have slowly moved

away from city centers (Jacobs, Ducruet, & De Langen, Integrating world cities into production

networks: The case of port cities, 2010). There are many reasons explaining this phenomena;

environmental awareness and the right for cleaner air (public good), lack of land to expand port

related business activities near city centers (Jacobs, Ducruet, & De Langen, Integrating world

cities into production networks: The case of port cities, 2010) (Hoyle, 1989) and governments

are more involved in spatial planning and designing creating specific places for these industries

to conduct their businesses. Furthermore, Jacobs (2007) argues that because port related

activities went from a more public (governmental) domain of business to private

(corporate/investors) business negatively impacting the port-city relationship because “the

dependence of ports on the urban labor market as well as the reduced dependence of cities on

ports for local economic growth” (Jacobs, Political Economy of Port Competition , 2007).

On the other side of production related businesses, we have services related activities. These

activities are mostly ICT, insurance and consultancy related businesses (see Jacobs 2007 et al.

table 1 for complete overview). It is tough to pinpoint that these sectors exist near ports regions

specifically to enhance/provide port business with for example consultancy or legal support. In

other words, it is near impossible to categorize financial service related businesses in port

regions that are there solely there because of the port simply because there isn’t extensive data

available to proof this.

3. THE HYPOTHESES

FDI is positively related to economic growth, more often than not regardless of a host country’s

human capital level. Furthermore, these investments can create jobs for in a region and increases

the competitiveness in the host country (Wang & Wong, 2009). As firms looking to make a

Foreign Direct Investment are assumed to be profit maximizers, they select an investment based

on the chosen region’s characteristics impacting profits relative to other regions (Makabenta,

2002). Some of the determinants of FDI inflows, based on existing theories are the market size,

the advanced infrastructure of big cities and ports, natural resources, the proximity to European

markets and finally, legislative and political risks (Ledyaeva, 2009). Accessibility to ports and

connected cities by rail and road are positively related to FDI locations especially for firms

seeking to locate in areas near ports to reduce transportation costs (Makabenta, 2002).

Makabenta (2002) shows in her research in the Philippines that the port and highway variables

have strong pull effects on manufacturing FDI, as potential investment areas do not only need

pools of skilled labour, but also need good transport options to markets and/or sources of

resources via ports and roads. The availability of a port, the marginal effect of the dummy

variable PORT, in the study increased the new FDIs by x1.62. This hints to the conclusion that

further improvement of the port (region) increases the attractiveness of foreign direct

investments. Similar results have been concluded in Nyamai & Wall’s (2015) research which

looked at competitiveness between port and non-port cities. The results showed that for the port

cities higher education and liner shipping were positive and significant indicating that smart

people and global shipping networks are needed to attract more FDI. Whereas for non-port cities

the employment rate was positive and significant indicating that an increase in employment

would increase FDI in the city. However, port cities remain the most competitive compared to

non-port ones because of the ability to attract FDI due to a growth in the number of smart human

capital (Nyamai & Wall, 2015).

Hypothesis 1: If a region includes a port, it is expected to attract more FDI compared to a

non-port region.

The growth of the port of Shanghai is related to the development of its economy, manufacturing

and foreign trade, this because Shanghai is a major exported of manufactured products.

Therefore, it is needed for this market to develop a major international port in/near Shanghai.

The port serves as a city-serving hub port-city, the port basically serves as an international

shipping center for the region for the mass-produced goods that need to be exported, as well as

the resources that need to be imported (Huang, 2009). To deal with these trades, a trade- and

financial service center will be set up in the city. This is supported by Makabenta (2002) as

shown in the results where manufacturing FDI firms looks for regions that have, among other

aspects, access to ports and highways as to which firms have to ability to efficiently transport

their manufactured products. These two effects also have the largest marginal effects on the

attraction of FDI. Furthermore, also European ports such as Rotterdam and Antwerp have

developed their ports to big facilities of production and manufacturing because they are

dependent on the import of raw materials and want to be close to the port (Jacobs, Ducruet, & De

Langen, Integrating world cities into production networks: The case of port cities, 2010).

Therefore, we would expect businesses within the manufacturing and “wholesale, warehousing,

logistics and transport” businesses to be drawn by the attractiveness of short distances or direct

accessibility of nearby ports due to the fact that the main goal of ports is to load and unload the

goods as quick and efficient as possible (Jacobs, Ducruet, & De Langen, Integrating world cities

into production networks: The case of port cities, 2010). Business sectors that might be drawn to

the port but need not to be near the port are those that are concerned with the service aspect of

the business.

Hypothesis 2: The effect of the port in attracting FDI is stronger for investments in the

Manufacturing sectors relative to the Services sectors.

Ledyaeva (2009) suggests that in general competition for FDI between regions with ports has

increased after the crisis. The negative spatial relationship in FDI within the group of port-

endowed regions indicates that if one region with a port can offer additional advantages in other

FDI determinants than what is offered by neighboring port regions, foreign investors will tend to

choose that region. Components of port competitiveness such as; cost‐related vessel and cargo

entering, efficient inland transport network, frequency of large container ships’ calling,

inland transportation cost, port accessibility/congestion/safety, professionals and skilled

labor in port operations and reliability of schedules in port are factors than can either

increase or decrease a port’s attractiveness (Yeo, Roe, & Dinwoodie, 2011)

A prime example of such port competitiveness are the ports of Shanghai and Ninbo. With the

ever-growing interest in FDI in China, the Yangtze River Delta is expanding exponentially. In

addition, two of the fastest growing container lines in the world, Cosco Container Lines and

China Shipping Container Lines, not only have their headquarters in Shanghai but also use

Shanghai as their main hub port in China. Competitive wise this is obviously better for the port

in Shanghai than Ningbo port. This is especially the case since Shanghai’s throughput is largely

domestic cargo, with international import and export still playing a major role within the port.

The potential for a further significant expansion in demand for the port of Shanghai is therefore

obvious (Cullinane, Teng, & Wang, 2006). However, it should be mentioned that in the future, as

the development of smaller ports such as Ningbo increases and with the ever-growing interest for

FDI in China, the port of Ningbo will benefit the greatest marginal benefit from the economies of

scale and efficiency improvement (Cullinane, Teng, & Wang, 2006). Especially because of the

ever-growing costs when dealing with larger ports, Hong Kong’s port charges for example are

the highest in the world and are at least 63 percent more expensive than other Asian ports (Yeo,

Roe, & Dinwoodie, 2011) In Asia, there appears to be a link between the size of the port and the

attraction of FDI to the port region. We will therefore test if this is the case in Europe as well.

Hypothesis 3: The size of the port is positively related with the number of total investments and

this relationship is stronger for large, main ports.

4. DATA & METHODOLOGY

4.1 Data

To examine the level of attractiveness of inward FDI between European regions explained by the

presence of a port, we chose the number of investment projects as the independent variable of

our analysis since it is a widely accepted operationalization of the competitiveness among the

multiple location alternatives in the economic research. Information concerning investment

projects were extracted from the FDI & Market dataset which includes the number of

investments per sector in 237 NUTS-2 regions for the period 2003-2011. To get information

about economic, demographic and structural characteristics on regional level, we merged the FDI

& Market dataset with the Regional Data Europe 2003-2010 dataset (Eurostat). Interestingly,

even though our main focus is on the impact of ports in regional economy, we could not find any

dataset to satisfy the needs of our research for those NUTS-2 regions and ports so we created one

of our own which perfectly includes all information needed for testing our hypotheses. More

specifically, our main motivation was to make a distinction between regions and non-port

regions based on the following concerns: regions with and without ports do not share apparently

same characteristics in terms of socio-economic infrastructure and institutions. They might

reveal differences between their main sources of financial income, the quality of human capital

or the levels of criminality and uncertainty. Not exploiting this heterogeneity on regional level

might lead to biased results and invalid conclusions. Segregating regions in two main

categories enables us to make comparisons between the levels of attractiveness in port and non-

port regions and isolate the effect of the presence of a port. Therefore, we created a dummy

variable which takes the value 1 if the region is recognized as a port region and 0 otherwise. This

variable characterizes 131 NUTS-2 regions and the procedure we followed to assign the proper

values was by looking at each region geographical position and considering it as port region if

this is directly connected to the sea and served by at least one port within or close to the borders

of that region.

Our unique FDI dataset

consists of totally 13,441

investment projects in 278

NUTS-2 regions across

European countries from 2003

to 2010. Investments,

however, are not normally

distributed as 75% of them

range from 0 to 9, indicating a strong

positive skewness. Almost 60% of the total

investments are in the Services sectors; Business, Financial, ICT, Transport and ConServ.

(Figure 1).

Regarding the business

activities, those can be

Figure 1: Distribution of FDI per Business sector

Figure 2: Distribution of FDI across Europe

viewed as upstream (R&D, headquarters) and downstream (Sales & Marketing, Business and

supporting services and Logistics). Most of the investments fall into the Sales & Marketing

category followed by production, indicating a market seeking behavior of the investing firms.

Finally, Figure 2 indicates that half of the total investments are found in Central and Western

Europe, while the UK and Ireland hold together 20% of them.

4.2 Ports and urban attractiveness

Capturing the impact of ports in attracting FDI, we set as variable of interest the dummy variable

‘portregion’. This measurement, on the one hand gives a relevant distinction between the

competitiveness of port and non-port regions as necessary to test the main hypothesis. However,

some regions might be located near the sea but are not be considered a “port region”. Whilst in

other cases, there may be large areas not near the sea that have a large port that are considered as

port regions.

Alternatively, we estimate this effect by using the distance to the nearest seaport. This

measurement overcomes some confusion derived from the different sizes of the regions. At the

same time, distance is a rather general indicator since we do not know exactly the points of

reference taken in the NUTS-2 regions. We expect, in consistence with our main hypothesis, a

positive coefficient for the port dummy and therefore a negative coefficient for the variable

distance to seaport.

4.3 Sectoral Analysis

The utterly different nature and

structure of MNCs in service

and manufacturing industries

Figure 3: Distribution of FDI in Manufacturing industries
per Business Activity

hint at different required circumstances and surroundings of the market of entry when it comes to

investing. Our interest turns to investigating possible differences between the effect of a port in

service and manufacturing industries. Unlikely, even our dataset provides information on the

number of investments through nine different sectors, the variation of the number of investments

between the sectors was not separately enough and exploitable. Thus, we summarized them in

two groups; ‘Manufacturing’ and ‘Services’. ‘Manufacturing’ consists of the sectors HighTech,

MedTech, LowTech and Proceeding Industry and ‘Services’ respectively, consists for Transport,

ICT, Consumer, Business and Financial Services. Since we do not know exactly if the

investments are related with port activities or not, we can’t have a clear disposition as for the

expected sign. We expect though a stronger effect for Manufacturing for reasons mentioned in

previous sections. A detailed breakdown on business categories, sectors, functions and activities

can be found in the Appendix. As mentioned before, it is expected the effect of the port to be

stronger for investments in Manufacturing. It is crucial here to have a closer look at the business

activities of these MNCs investing in manufacturing. As seen from Figure 3 only one third of

these investments are related to production plants in contrast with upstream and downstream

activities, something useful to keep in mind when it comes to interpret our results.

4.4 Effect of Port size

Testing our third hypothesis requires the classification by size of the ports included in our own

created dataset which ensued from the following procedure: first we matched each of the port

regions with one -the largest and busiest in the area- port. In order to classify them, we utilized

the list of the twenty largest ports in Europe by the volume of TEUs in 2010 by Eurostat. For the

rest of the ports, we searched the numbers of TEUs by looking at relevant articles and annual

reports or websites of the port itself. Twenty-Foot Equivalent Units (TEUs) are counted annually

and individually for all ports and functions as a measurement of the largeness or busyness of a

port. Note here that for 11 out of 101 ports chosen, data on TEU volumes were not available.

Thus, we got a general picture of their size by looking at the volume of cargo they handle in

tons-from the same sources and by comparing them with the numbers of the ports. We

subsequently assigned each port properly in our ranking list. We therefore categorized ports in

four groups: Small ports (0-100.000 TEU), Medium ports (100.000-300.000 TEU), Large ports

(300.000-1.000.000 TEU) and Mainports (1.000.000 – more TEU) where ‘Small Port’ is selected

to be the bench category. The distribution of these ports can be found in Table 6 and Table 7 of

the Appendix.

Based on this and our theoretical background about the contribution of ports to the urban

competitiveness, we expect the effect of the presence of a port to the attractiveness of inward

FDI to be positive and stronger for main, large ports that act as hubs in the global networks.

For reasons mentioned in the previous section, we also expect that port size will be also

significant through different sectors, with stronger to be that on manufacturing.

4.5 Control Variables

Following the taxonomy of FDI motives as reported by Serwicka, Jones and Wren (2014) which

was based on Dunning’s OLI framework, we introduce our control variables by linking them to

the main motivations of inward FDI, adjusted to the European environment. Here, diverse

motives are translated as market-, resource-, efficiency- and strategic-asset-seeking FDI.

Market-seeking FDI aims to serve overseas demand and is mostly driven by factors like the size

of the host country and market and its singularity compared to the neighboring ones. We use

gross domestic product (GDP) per capita, as a measurement of the regional economy, widely

used in the investment literature and population in thousands of residents as a proxy for the size

of the country. We also include population density to control for unequal distribution of

population among the regions. The percentage of population with higher education following

the ISCEC classification by Eurostat in 2011, is included to capture the demand for high-skilled

employees which is mostly desired in-service industries and business upstream activities.

Resource and efficiency-seeking FDI aim at the best quality of production with the lowest

possible costs. In other words, MNSs tend to maximize their benefits by exploiting the

differences between the costs of acquiring these resources of production in the mother and the

host country. To capture labor costs, we include the average of annual wages per region.

Despite numerous previous studies, we do not include long-term unemployment in our

estimation because its effect is rather ambiguous with both negative and positive impacts; high

unemployment makes recruitment easier and cheaper in theory but it may also induce serious

socio-economic problems like rising crime and skills extinction, creating this way uncertainty

and unattractive surroundings for the investors. Another cost taken under consideration is

corporate taxes as a proxy for capital costs which differs to a great extent between the European

countries. To illustrate with an example from our sample, corporate taxes range from 9%

(Cyprus) to 40% (Germany). While the previously mentioned types of motives exploit actually

existing assets of the regional economy, strategic-asset-seeking FDI aims deeper, to acquire

foreign assets of the economy of entry. R&D is of immense importance for firms when strive to

create competitive advantages over their competitors. It enables companies to predict and –

hopefully- meet future demand and trends and be ahead of fellow companies within their sector.

In this sense, expenses on R&D should be better seen as a kind of investment and not as an

expense per se. Based on the definition by OECD on gross domestic spending on R&D3, we

incorporate in our model such expenses to capture the difference in the levels of attractiveness

between the alternative locations for those MNEs which work toward establishing well-founded

and profitable operations in a foreign market.

Numerous empirical studies have verified that MNCs are attracted by the positive externalities of

economic concentrations within both the same and different sectors (Bronzini, 2004).Marshall in

his book ‘Principals of Economic Theory’ (1890) made a distinction between localization and

urbanization economies which are driven by lower input cost, larger labor markets and

knowledge spillovers. It would be absurd to assume that investments occur randomly through

space and time and independently to each other so not controlling for economic concentrations,

might lead to bias estimates and imprecise conclusions. In order to capture the effect of

localization and urbanization we include in our model the share of the population of own

employment in 5 sectors and the percentage of urban land use respectively. In the first case the

firm has to benefit from lower input costs and specialization in each ever field and in the second,

from the amenities of a big area like in transportation and infrastructure regardless the industry.

4.6 Estimation Strategy

MNCs location choices indicate the attractiveness of each region and subsequently its

competency to pull in FDI, able to elevate the regional economic status. The nature of our

dependent variable commands the application of count data models where counts are the number

of events occurred within a fixed time period. In our attempt to specify the probability of an

3 Gross domestic spending on R&D is the total expenditure (current and capital) on R&D carried out by resident

companies, institutes, universities etc., in a country. It includes R&D funded from abroad, but excludes domestic

funds for R&D performed outside the domestic economy. This indicator is measured as a percentage of GDP.

investment to occur in a particular region, OLS would yield biased estimates so we turn to a

Poisson process. Poisson however, is violated more often than not since it is based on the

equality of conditional mean and variance. From descriptive statistics, we see that the variance of

our dependent variable is more than six times its conditional mean, a clear sign of

overdispersion. We first calculate a Poisson regression-even though we believe that the Poisson

distribution is not the appropriate one- in order to test the goodness of fit of the model. The chi

square statistic is very significant in this case; thus, Poisson is not the best option. To continue, a

generalized version of Poisson is implemented, the negative binomial model, which has an extra

parameter that controls for extra variation. The likelihood ratio test for over dispersion confirms

once again our selection for the negative binomial model. As discussed, we incorporate along

with the variables of interest, typical controls aiming to inferring about the main determinants of

investing in Europe.

5. RESULTS

The results of our negative binomial regression model are represented in Table 1. The most

important findings are depicted in the two first rows, “portregion” and “distancetoseaport”. To

test the first hypothesis, the number of investments is taken as the independent variable. As seen

from the estimation of the baseline Model 1, the portregion dummy is highly significant and

positive, clearly demonstrating that port regions attract more FDI compared to non-port regions.

The results of the alternative estimation by using ‘distancetoseaport’ in column two, are

consistent with previous findings indicating that if the distance from a region to seaport

increases, the number of investments in this region decreases. Accordingly, we find strong

support for hypothesis 1.

Columns 3 to 6 show the estimations of our second hypothesis which concerns the effect of the

port in different business sectors. With the number of investments in both sectors taken as the

independent variable, colums 3 and 4 reveal that the port has a strong effect in attracting

investments of both Manufacturing and Services industries. Contradictory with our expectations,

is the stronger effect for investments in Services. This result is well-based on extant literature

and will be discussed in our conclusions.

The consistent results of the alternative estimation are shown in columns 5 and 6. As both

coefficients are significantly negative, a larger distance to the nearest seaport implies a lower

number of investments for both sectors. This effect is again stronger for Services indicating no

support for our second hypothesis.

Tabel 1: Port Regions and Sectoral Dissimilarity

Findings regarding the effect of the port size in attracting FDI are represented in Table 2. The

baseline estimation depicted in column 1, offers partial support to our third hypothesis, yielding

an insignificant coefficient for the third size category ‘Large ports’. The significant, positive and

large coefficient in the fourth category ‘Mainports’ indicates that the expected log number of

investments in regions with a major port is 1.819 higher compared with this number in regions

with very small ports. As for ‘Medium Ports’ the result is significant at a 1% level, has the

expected positive sign and signify that the expected log number of investments in regions with a

port of category 2 is 0.665 higher than regions with a port in category 1.

Since data are available, we proceed with a sectoral analysis to control for possible differences in

the port size effect between attracting FDI in Manufacturing and Services Industries. Once again,

Services sectors seem to be more responsive in the presence of a port in an area. What is

paradoxical here, is the insignificant coefficient of the ‘Mainport’ category in investments in

Manufacturing. Extremely high standard errors and relatively big effects could be signs of

inadequate variation within the business sectors or omitted variable bias.

The results of the rest independent variables are identical through all the estimation models and

yield some critical findings about the determinants of investing in European regions for the

period 2003-2010. Those are discussed in the next section.

Table 2: The Port Size Effect

6. DISCUSSION & CONCLUSION

Our findings on the attracting effect of ports are in line with the idea that some cities and ports

are favored by nature, being in proximity to the main lines that connect different parts of the

world within the global networks (Zhao, Xu, Wall, & Stavropoulos, 2017). Ports play a

significant, positive role in the ability of urban competitiveness to attract FDI. Examples of such

ports and cities in our sample are the Mainport of Rotterdam, the port of La Havre in France and

the port of Felixstowe in UK. The NUTS-2 classification is particularly useful in socio-economic

analysis of regional policies4. In our case this acts as a limitation since it provides information on

aggregated areas. It is expected that the port effect will vary across different areas in a NUTS-2

region basically due to the different sizes and characteristics of these regions. To eliminate that

bias we estimated in two ways our model and the results were consistent. An interesting aspect

would be to make a case based on city and not regional information in order to eliminate any bias

induced by geographical and economic clusters. That would yield useful information for policies

and institutional strategic planning that strive to escalate the urban economic performance

influenced by the existence of a port in this particular area.

The fact that the effect for investments in Services is stronger than for Manufacturing, might not

satisfy our expectations, nonetheless there are some arguable explanations on this. As mentioned

in ‘Introduction’, Manufacturing sectors tend to concentrate in Eastern Europe. Clearly, this is

not the case here as main European ports are in the Western part of the continent. This finding

confirms the stronger effect we had for the Service Sectors but is not enough to rest in our

laurels, so we also considered some additional reasons. Extant literature suggests that port cities

have been developed far and segregated from their ports because of the rise of advanced service

4 http://ec.europa.eu/eurostat/web/nuts

http://ec.europa.eu/eurostat/web/nuts

industries where the port does not meet the requirements for such operations (Zhao, Xu, Wall, &

Stavropoulos, 2017). At the same time, ports create excellent circumstances for companies

specialized in advanced producer services, especially for maritime and port-related services

(Jacobs, Political Economy of Port Competition , 2007). It is also important to take under

account that regions with most investments in our sample are big financial centers which

facilitate international firms in shipping and maritime activities like insurances and other

supportive services such as London and Amsterdam. A list if the top 10 NUTS-2 regions with

their main specialization can be found in Table 3 in the Appendix where only three of the 131

port regions defined in our study are included. If with past data the effect for FDI in services was

stronger, imagine how different this effect could be today after 8 years of incessant advance and

automation integration. It might be also the case that investing MNCs have entered a particular

region driven by the attractiveness and the status of the host country where the latter, in turn, can

enhance the general performance of the former (Karreman & Van der Knaap, 2010). One

direction for further research is to collect adequate data that capture all these port and urban

characteristics and infer what matters more for location choices; the port, the city or a

combination of those two? Highly influential port-specific factors like the port authority – private

or public – should be taken under account as well as urban-specific characteristics like the

potentiality of a market captured by growth rates. In our case it was impossible due to data

unavailability but we would be glad to incorporate such aspects in a future study.

Finally, we find a positive relationship between the log number of total investments and presence

of a main port in a region. The confounding result on the sectoral heterogeneity has three

possible explanations. First, the port size is not the most important criterion. Commercial and

geopolitical conditions as well as just the big number of ports in an area are factors able to set the

port in the center of the urban economic activity. (Roa, Peña, Amante, & Goretti, 2013).Second,

In the same study, it is found that large ports do not exhibit high potentiality in goods

management but instead features that downplay its importance as for the volumes handled and

the size of the hinterland. Third, the ports included in our study are mainly deep-water seaports

that basically handle freight (measured in TEUs), neglecting recreational, military, fishing and

tourism activities as well as river, harbor and terminal ports. In both cases, a thorough analysis

must be done distinguishing between the different types of port and activities in order to account

for this heterogeneity and draw specific conclusions, appropriate for inference to a greater

population.

As it seems, FDI in Europian regions is strongly driven by efficiency and strategic-asset seeking

motives. Investing MNCs strive to benefit from costs differences between the country of origin

and entrance and is becomes apparent from the significance levels (1%) of the corresponding

controls -wages, taxes, R&D expediture. The latter especially is in line with what Dachs et al.

stated that R&D has been overly extended in Europe and is inescapable for multinationals that

strive to develop goods and services outside their country and not only produce and sell them

(Dachs, Kampik, Scherngell, & Zahradnik, 2012). At first sight, own sector employment does

not display any important role in attracting FDI, something hard to accept as true. However, high

standard errors combined with high standard deviation obtain by summary and descriptive

statistics, may indicate the following: our sample is not really representative of the actual

population or sector specialization varies a lot in our sample and our model in not appropriate to

exploit all this explanatory power that these variables add in the model. In every case, we give a

suggestion to any future researcher whose focus will be on the contribution of ports in the

attractiveness and socio-economic performance of the cities: it is impossible to draw generalized

conclusions not taking under account different port types and activities or controlling for cities

heterogeneity, so it might be more informative to make comparisons between the different

regions and cities of the same country rather on a continent level.

REFERENCES
Bevan, A. A., & Estrin, S. (2004). The determinants of foreign direct investment into European transition

economies. Journal of Competitive Economics, 775-787.
Bronzini, R. (2004). Foreign Direct Investment and Agglomeration: Evidence from Italy. European

Regional Science Association.

Burger, M. J., Van der Knaap, B., & Wald, R. (n.d.). Revealed Competition for Greenfield Investments

between European Regions. Tinbergen Institute.

Cullinane, K., Teng, Y., & Wang, T. (2006). Port competition between Shanghai and Ningbo. Maritime

Policy & Management.

Dachs, B., Kampik, F., Scherngell, T., & Zahradnik, G. (2012). Internationalisation of business

investments in R&D. Austrian Institute of Technology.

Disdier, A.-C., & Mayer, T. (2004). How different is Eastern Europe? Structure and determinants of

location choices by French firms in Eastern and Western Europe. Journal of Comparative

Economics, 280-296.

Dooms, M., Lugt, L. v., & Langen, P. W. (2013). International strategies of port authorities: The case of

port of Rotterdam Authority. Research in Transportation Business & Management.

Ducruet, C., Rozenblat, C., & Zaidi, F. (2010). Ports in multi-level maritime networks:evidence from the

Atlantic. Journal of Transport Geography.

Dunning, J. (1981). International Production and the Multinational Enterprise. London: Allen & Unwin.

Dunning, J. H. (2001). The Eclectic (OLI) Paradigm of International Production: Past, Present and Future.

International Journal of the Economics of Business, 173-190.

Ferrari, C., O., M., Bottaso, A., & Tei, A. (2012). Ports and Regional Development: A European

Perspective. Paris: OECD Regional Development Working Paper.

Ferrari, C., Parola, F., & Gattorna, E. (2011). Measuring the quality of port hinterland accessibility: The

Ligurian case.

Transport Policy.

Hall, P., & Jacobs, W. (2012). Why are maritime ports (still) urban, and why should policy-makers care?

Maritime Policy Management.

Hoyle, B. (1989). The port—City interface: Trends, problems and examples. Geoforum, 429-435.

Huang, Y. (2009). The growth of global hub port cities under globalisation: The case of Shanghai

international shipping centre.

Jacobs, W. (2007). Political Economy of Port Competition . Nijmegen: Academic Press Europe.

Jacobs, W., Ducruet, C., & De Langen, P. (2010). Integrating world cities into production networks: The

case of port cities. Global Networks, Wiley, pp.92-113. .

Karlsen, T., Silseth, P. R., Benito, G. R., & Welch, L. S. (2003). Knowledge, internationalization of the

firm, and inward–outward connections. Industrial marketing management.

Karreman, B., & Van der Knaap, B. (2010). The Geography of Equity Listing and Financial Centre

Competition in Mainland China and Hong Kong. Journal of Economic Geography.

Kolstad, I., & Wiig, A. (2012). What determines Chinese outward FDI? Journal of World Business.

Kostiainen, J. (2002). Urban economic development policy in the network society. Tampereen Yliopisto.

Ledyaeva, S. (2009). Spatial Econometric Analysis of Foreign Direct Investment Determinants in Russian

Regions. The World Economy.

Lovering, J. (2003). MNCs and wannabes: inward investment, discourses of regional development and

the regional service class. Cheltenham, UK: Edward Elgar.

Makabenta, P. M. (2002). FDI LOCATION AND SPECIAL ECONOMIC ZONES. RURDS Vol. 14.

Marshall, A. (1890). Principles of Economics ,8th edition.

Merk, O. (2014). The Competitiveness of GlobalPort-Cities:Synthesis Report.

Merk, O., & Hesse, M. (2012). The Competitiveness of Global Port-Cities: The Case of Hamburg,

Germany . OECD Regional Development Working Papers.

Meyer, K. (1995). Foreign direct investment in the early years of economic transition: a survey.

Economics of Transition, 301-320.

Nijdam, M., & van der Horst, M. (2018). Port definition, concepts and the role of ports in supply chains.

In Ports and Networks (p. 10).

Abingdon: Routledge.

Notteboom, E. (2010). Concentration and the formation of multi-port gateway regions in the European

container port system: an update. Transport Geography, 567-583.

Nyamai, D., & Wall, R. (2015). WEALTH AND THE CITY: The Competitiveness of Port cities and Non-

port cities. Rotterdam.

Porter, M. E. (2000). Location, Competition, and Economic Development: Local Clusters in a Global

Economy. ECONOMIC DEVELOPMENT QUARTERLY, 15-34.

Roa, I., Peña, Y., Amante, B., & Goretti, M. (2013). Ports: definition and study of types, sizes and

business models. Journal of Industrial Engineering and Management, 1055-1064.

Sáez, L., & Periáñez, I. (2015). . Benchmarking urban competitiveness in Europe to attract investment.

Cities.

Song, L., & van Geenhuizen, M. (2014). . Port infrastructure investment and regional economic.

Transport Policy.

Stopford, M. (2009). Maritime Economics . In M. Stopford, Maritime Economics (3rd ed., p. 81).

Abingdon: Routledge.

Wang, M., & Wong, S. (2009). What Drives Economic Growth? The Case of Cross. Kyklos.

Yeo, G., Roe, M., & Dinwoodie, J. (2011). Measuring the competitiveness of container ports:

logisticians’ perspectives. European Journal of Marketing.

Zhao, Q., Xu, H., Wall, R., & Stavropoulos, S. (2017). Building a bridge between port and city:

Improving the urban competitiveness of port cities. Journal of Transport Geography, 120-133.

APPENDIX

Table 3: Investment Portfolios of the Top 10 NUTS-2 Regions

Table 4: Business Sectors Taxonomy

Table 5: Business Functions Taxonomy

Table 6: Distribution of ‘Large’ and ‘Main’ ports in TEU

0 2000000 4000000 6000000 8000000 10000000 12000000 14000000

Port of Rotterdam

Port of Antwerp

Port of Hamburg

Port of Bremen

Port of Algeciras

Port of Valencia

Port of Felixstowe

Port of Gioia Tauro

Port of Piraeus

Port of Marsaxlokk

Port of le Havre

Port of Genoa

Port of Brugge-Zeebrugge

Port of Southampton

Port of Barcelona

Port of London

Port of Las Palmas

Port of Marseille

Port of Gdansk

Port of Gothenburg

Port of Livorno

Port of Venice

Port of Liverpool

Port of Bilbao

Port of Dublin

Port of Lisbon

Port of Napoli

Port of Leixoes

Port of Sines

Port of Helsinki

Port of Aarhus

Port of Trieste

Port of Klaipeda

Port of Teesport

Port of Thessaloniki

Port of Grimsby

Port of Milford

Port of Dunkirk

Port of Riga

Port of Limassol

Zeeland Seaports

Distribution of ‘Large’ and ‘Main’ ports in TEU

Table 7: Distribution of ‘Small’ and ‘Medium’ ports in TEU

0 50000 100000 150000 200000 250000

Port of Belfast
Port of Grangemouth

Port of Vigo
Port of Ravenna
Port of Messina
Port of Rostock

Port of Tallinn
Port of Oslo

Port of Nantes
Port of Ancona

Port of Heraklion
Port of Rochelle

Port of Copenhagen/Malmo
Port of Bristol

Port of Cartagena
Port of Szczecin

Cromarty Firth Port
Port of Ghent

Port of Plymouth
Port of Sandefjord
Port of Stavanger

Clydeport
Port of Civitavecchia

Port of Cardiff
Port of Gijón

Groningen Seaports
Port of Bordeaux

Port of Odense
Port of Palma de Mallorca

Port of Stockholm
Port of Tyne

Port of Amsterdam
Port of Bergen

Port de Saint Valery sur Somme
Port of Barrow

Port of Corfu
Port of Kavala

Port of Koge
Port of Monaco

Port of Oskarshamn
Port of Patras

Port of Trondheim
Port of Melilla

Port of Brest
Port of Aalborg

Port of Aberdeen
Port of Faro

Port of Galway
Port of Pescara
Port of Tromsø

Port of Santander
Port of Brindisi

Distribution of ‘Small’ and ‘Medium’ ports in
TEU

Table 7: Distribution of ‘Small’ and ‘Medium’ ports per TEU

Abstract

This paper focuses on the question whether and to what the presence of a port affects the

greenfield FDI in the European NUTS-2 regions. The hypotheses have been derived from

existing international business and economic geography research context. In this paper

truncated negative binomial model and random effects model have been performed as

empirical methods of analysis and data regarding greenfield FDI from fDi Markets and freight

from Eurostat for the period 2003 – 2010 have been used. The findings of this study show that

there is no or little evidence that the presence of a port in a European region lead to more

greenfield investments in the same region.

1. Introduction

Jungnickel (1993) proclaims the most noticeable growth in trade took place in the 1970s.

Subsequently, the augmentation in foreign direct investment (FDI) has gone beyond of both

– national outputs and international trade. The second half of 1980s is noted with 33 per cent

annual growth rate of FDI (UNCTAD, 1993). Over short period of time the world share of the

latter has doubled. As remarkable as such event may be, the increase and distribution of FDI

can be accounted for only an insignificant portion of all international capital flows coursing

through highly integrated world financial markets (Banuri and Schor, 1992).

More than 20 years ago Dunning (1993) defined the “ownership–localization–internalization

(OLI) paradigm” which outlined the activities of multinational enterprises (MNE’s) as the ones

derived from the relationship between the competitive advantage of MNE’s and countries

and their ability to generate value-added activities. Firms prefer to invest abroad when

ownership advantages can be obtained by capitalizing possessory assets/capabilities and by

profiting from the location-specific advantages rendered from internalizing cross-border

activities in receiving country. Setting apart the definition of horizontal and vertical FDI’s,

Dunning (1993, 1998) specified four key determinants for the companies to internationalise

the operations and emphasize the specific-location advantage: (1) the natural resource –

seeking incentive (admission to natural resources); (2) the market – seeking incentive

(admission to new markets); (3) the efficiency – seeking incentive (reorganization of

operations in order to diminish the labour, machinery and material costs of production and

increase efficiency); (4) the strategic – asset – seeking incentive (admission to strategically

related created assets). Both, international business and economic geography research, focus

their interest on the notion of location – where and why the companies opt for explicit

destination of their activities (e.g. Nachum & Wymbs, 2005; Porter, 2001; Krugman, 1991a;

Markusen, 1996). The success of any location, would it be a country, a region or a city

predominantly relies on its attractiveness compared to other locations in terms of availability

of local resources (Robinson, 2002). The ports have traditionally served as means to connect

domestic and international markets and impact the trade.

The ports in their essence contribute to the value of the shippers and to the third-party service

providers; the value projection is based on the consumer segmentation and targeting, thus

facilitating the ports to obtain value not only for themselves, but also for the chain those are

implanted in. As the status of ports changes from sites with distinct functions to elements of

value-driven chain systems, a paradigm shift takes place (Robinson, 2002). The role of the

ports as a transportation or production hubs has been greatly neglected by the academics

focusing on the world-wide commodity or production chains, regardless their contribution to

the global allocation of goods (vessels carry about 90 per cent of world trade volumes) (Jacobs

et al., 2010).

This paper is dedicated to the analysis of greenfield FDI, focusing on the ports of European

NUTS-2 regions during the period 2003 – 2010. The main concern is whether and to what the

presence of a port affects the FDI in European regions. The paper is organized as follows: in

Section 2, the focus is given to the theoretical framework and the presentation of the

hypothesis. Section 3 provides introduction of data and empirical methods used for analysis.

Section 4 continues with the discussion of the latter and the results obtained. Section

contains the conclusions and suggestions for the future research topics.

2. Literature review

2.1. Background

To what extent the ports keep having an influence on the hosting city is a question of general

concern. Their relationship and interdependence seem to be decreasing both in Europe and

in Asia, where the comparison has been drawn and different reasoning for each continent

concluded (Ducruet, 2006). The costs of conducting a port activity, accompanied by higher

levels of pollution and congestion, are increasing and since transportation costs decrease,

negative externalities expand into the hinterland (Hall, 2007). Both, costs and profits, are

spread over a wider region, immersing into the latter.

Nowadays, a new stream of port development has appeared – ‘port regionalization’ – due to

the growing importance of logistics networks and emerging locations outside the traditional

port territory. Within the port regionalization concept, the activities of the port are partially

relocated into the hinterland, where opportunities to improve capacity and freight circulation

due to affordable and extensive spatial capabilities arise. Such process is predominantly

market-driven and requires high levels of efficiency gains (Notteboom et al., 2005). Although,

the topic of port regionalization has not been well-studied from a spatial econometric

perspective, Bottasso et al. (2014) found that ports possess tendency to increase GDP of the

region where located and have positive effect on the GDP of neighbouring regions.

Ballou (2007) in his research has indicated that under the feasible hypothesis regarding a

general trend of increased globalisation and free trade in combination with increasing

outsourcing activity, the logistics activity will gain more importance and growth, which in turn

indicates an inflow of FDI due to the internationalization of the economy. Thus, it is not

surprising that the region of Lombardy (Italy) received 43% of inward FDI of the Italian logistics

sector in 2010 (Maggi et al., 2011), while it does not have a direct access to the nearest

seaport (located in the region Liguria) and Milan, the regional capital, is the main economic

agglomeration of the region Lombardy.

The topic of spatial distribution of the economic activity has been well-studied prior. The

neoclassical location theory presupposes the idea that the market will converge to an

equilibrium. In the long-term, the economic activity will locate ubiquitously through the

territory due to the constant returns to scales. Hence, the more productive industries will

locate in the core regions, where costs are higher. In the early 1990s a theoretical framework

of “New Economic Geography (NEG) models” has emerged (e.g. Krugman, 1991b; Puga,

2002). These models consider the chance that diversifications in the accessibility and

attractiveness of the locations caused by FDI into infrastructure could possible contrary

impact on the spatial distribution of economic affairs. The models vary from the preceding

ones because of the market seeking (market size) and cost saving (production cost)

imbalances. As such, this indicates convergence of economic affairs also encouraged by

trading activities of the regions. Thus, agglomeration economies, affected by firms’ location

decisions, become endogenous (Ottaviano, 2008). Additionally, locations in a proximity to

gates and hubs are favoured due to efficient transport infrastructure (Krugman, 1993).

Traditionally, economic theory has been proposing that transport infrastructure positively

affects industry productivity predominantly by diminishing transport and time costs, leading

to decrease of production costs, increase in productivity, specialization, trade and

competition advancement, expansion of compatible markets (Aschauer, 1989); better

accessibility and low transportation costs assist in escalation of market capabilities of various

places (Niebuhr, 2006; Condeco-Melhorado et al., 2011). The local and regional government

officials are known to have supporting attitude towards transport infrastructure (where ports

own one of key roles) as it is a determining component to promote territorial cohesion,

economic development and concentration, moderate economic inequalities (Notteboom et

al., 2005). For example, for the period 2007 – 2013 around 30% of the EU Regional

Development Funds (ERDF) and Cohesion Funds have been dedicated to organize investments

with transport infrastructure (intended at enhancement of regional benefits and finalizing

trans-European Transport networks (TEN’s)) receiving a notable portion of the budget

(ESPON, 2009).

2.2. Hypothesis

Port as a magnet for regional FDI in Europe

Nowadays, MNEs more and more regard the continent Europe as a moderately unified

territory instead of a group of independent countries. Thus, regions in Europe with analogous

features located in different countries are frequently recognized as comparable replacements

than unalike regions of the same country (Basile et al., 2009). The regional competition has

increased due to the process of European integration, which has (1) shaded national

boundaries; (2) steadily eliminated economic, social and cultural dissimilarities among the

countries; (3) promoted free movement (capital, goods and workers )(e.g. Cheshire and

Gordon, 1995; Budd, 1998; Gordon, 1999; Lever, 1999; Markusen and Nesse, 2007).

Consequently, the groups embodying economical concerns of the area are competing, not

the regions themselves (Gordon and Cheshire, 1998). The local and regional government

officials under stress of the elections participate in the competitive ventures to impact the

employment and protect business premiums. Besides that, the officials desire to be

recognized for their dynamic role in exhilarating local and regional economic development

(Turok, 2004; Markusen and Nesse, 2007). The modern zeals to lure the FDIs in the region

taken by the officials include enhancement and application of regional assets associated with

specific labour pools, university research and societal background (Raines, 2003; Turok, 2004).

The stimulus is focused on the affinity of inward investments by powering emphasis and

support of the peculiar robustness of the location (Raines, 2003).

The regional trade (i.e. economic) activities significantly rely on the accessibility of port

facilities and services. As many industry experts and researchers note, one of the primary

instruments to develop port and its’ hinterland is through FDI inflow. Those can positively

impact economic health of the regions as well as of the country (Cho and Ha, 2009). As the

previous capital instalments of the investees (in contrast to the purchase of the present firms)

do not impact the location choices related to greenfield investments, the latter are beneficial

in the research of regional competition (Burger et al., 2012). Thus, the presence of the port in

a region can have a positive effect on the FDI inflow of the same region.

H 1: There is a positive relationship between the presence of the port in the European

region and the number of greenfield investments in the same region

Port, World City and regional FDI in Europe

Megacities and global cities are two different concepts that can be wrongly interchanged. The

first one is just a big agglomeration of people, such as Calcuta, in India, or Chongqing, in China.

What make them different from a world city is that both are missing cosmopolitanism and

interconnectedness environment (Goerzen et al., 2013). Sassen’s (1991) concept of “Global

capacity” inspired a hierarchical theoretical classification of 10 alphas, 10 bettas, 35 gammas

and 67 delta global cities (Beaverstock et al., 1999). Sassen (2000) concluded that the

globalization of capital was bringing to such cities as Tokyo, Singapore and London a

significant share of corporate powers. These cities are taking, for instance, the power of the

international finance (Beaverstock et al., 1996). Additionally, these cities are also strategic

locations for the internationalization of the advance product services (APS) firms, due to the

high educated labour pool and for being close to the clients (Bagchi-Sen, 1997).

As prior mentioned, Dunning (1998) has established a link between FDI flow and MNE’s

location strategies. Hence, the companies desire to locate in places maximizing corporate

objectives. A global city alone may be limited to receiving investments from one specific

service/industry sector, while port city may be the goal destination of the other, as both

locations show different characteristics. However, global and port cities together would have

a greater range of characteristics and options to offer, with different and complementary

labour pools, which could attract more firms from different sectors whose aims are

completely diverse.

H 2: The positive relationship between the port in the European region and the number of

greenfield investments is stronger in the presence of a world city in the same region

Port, Infrastructure and regional FDI in Europe

Ports have arisen as important places for logistics since supply chain management has

become one of the key components of the competitive advantage for firms and companies

(Li et al., 2006), as important places of trade and exchange of transport modes for

commodities and manufactures. As it has been said, ports are important for the management

of physical flows, transporting and storing, and by doing so, ports need physical infrastructure

connected with the city and the region. Transport infrastructures (defined by Rietveld (1994)

as “facilities such as railway lines and stations, highways, canals, sea- and airports”) are known

to be useful to speed up the convergence process between regions and economic growth. For

instance, Démurger (2001) found that variation between Chinese provinces’ economic

performance could be partly explained by transport infrastructure and telecommunication

facilities, while Boopen (2006) performed an analysis in Africa, providing also positive

relationship. But to take advantage of the port, both systems should be complementary and

integrated, and ports should have good connectivity to spread their economic potential.

The ports can be classified, according to their transhipment incidence, as: pure transhipment

hub (where containers are stored temporarily after being handled), hub (where transhipment

is combined with export-import activity), regional gateway (where the presence of

transhipment is lower than the export-import activity), and gateway/feeder (mostly focused

in export-import activity). The first kind of ports do not to have a great economic impact on

the surrounding areas (Musso et al., 2004), examples of these are Gioia Tauro and Algeciras,

located in Calabria (Italy) and Cadiz (Spain), two of the most depressed regions in Europe

where the percentage of transhipments is higher than 90%. On the other extreme, gateway

and feeder ports have a lower percentage of transhipments, for example the cases of the

biggest European ports (Rotterdam, Hamburg and Antwerp) the transhipment ranges

between 29 and 36%, so they must have good connections with hinterlands, good examples

are Rotterdam (The Netherlands) and Los Angeles (USA) (Rodrigue et al., 2013). In such

circumstances, ports may act as engines for the companies located close by or may attract

firms to the region, as those can make use of the offered infrastructure. Nevertheless,

congestion of transportation infrastructure may have negative effect on the total travel time,

also known as the “Braess paradox” and suggests that a rise in the supply of infrastructure

has a negative impact on the productivity (Sheffi, 1985). Thus, to maximize the profits the

port needs to be well connected (in combination with manageable traffic situations) and

integrated in a regional and national network. In the cases when the connectivity by road,

train or air is better, the economic impact would be higher and thus the port would attract

more FDI.

H 3: The positive relationship between the port in the European region and the number of

greenfield investments is stronger in the presence of good infrastructure in the same region

Port, Business Activity and regional FDI in Europe

Ports have been important places to relocate physical goods from one transport mode to

another for international trade. These factors used to facilitate investment and urbanization

of cities. However, ports have become irrelevant for the prosperity of cities (Fujita and Mori,

1996), as negative externalities associated to them, such as congestion, have arisen,

weakening the relationship between ports and port cities (Zhao et al., 2017).

Since the wave of national deregulations started in the 1970s and the proceeding

globalisation, there has been in a process of industrial relocation which has been studied by

two streams of analysis, called World Network Cities (WNC) and Global Commodity Chain

(GCC). While the first ones are analysis of the networks of information and advanced product

services (APS), this is, financial companies and law firms; the second are the analysis of the

production units connected through flows of commodities, where ports have an important

role. These flows seem to have an uncertain relationship (Jacobs et al., 2010), as APS tend to

follow the urban hierarchy. Cases about this can be found in Canada (Slack, 1989) and

Australia (O’Connor, 1989). Three main types of locations have been identified by O’Connor

(1989): firstly, port cities providing basic daily services; secondly, port cities managing long-

term contracts and thirdly, international cities managing the overall worldwide shipping

industry. On the other hand, ports are becoming important in the process of the supply chain,

thus they have the necessity to provide logistic facilities (Pettit and Beresford, 2009). This can

lead to expectation that surrounding regions of a port may receive FDI from companies that

take part of supply chains, which are companies linked between each other to add value in a

product that should be delivered to a customer (Christopher, 1992). Thus, manufacturing,

transport and logistic firms may be expected to locate in these port regions to take advantage

of the facilities of the seaport. The clustering of such companies then may attract FDI from

the same sectors and industries. On the contrary, APS firms will locate in global cities

specialized in information sectors that don’t need to be close to the physical flow of goods.

H 4: The positive relationship between the port in the European region and the number of

greenfield investments is stronger for the logistic sector projects of the same region

3. Data & Methodology

3.1. Data and Variables

The primary data sources that have been used, are the fDi Markets database and Eurostat

database for the timeframe 2003 – 2010, both on NUTS-2 level. The fDi Markets database

(26.995 observations) contains data on project level about the greenfield investments

between different sectors/business activities and the presence of different world cities. The

spatial distribution of total number of investments across two hundred sixty NUTS-2 European

regions (EU-25, Switzerland, Norway, Cyprus) is provided in Figure I. The investments are not

evenly distributed, it can be noted that regions including the world cities receive greater share

of FDIs.

Figure I – Total number of investments across 260 NUTS-2 European regions

The Eurostat dataset (2.120 observations) contains data about regional specific effects, such

as GDP per capita, long-term unemployment rate, percentage of people with high education,

accessibility, etc. Another dataset contains data (1.101 observations) about the amount of

freight (in thousand tons) on NUTS-2 level and is also coming from Eurostat. The importance

of this data is justified due to the reason that it shows in which region maritime transport

activity is present and thus indicating the existence of the ports.

To make the data suitable for the empirical analysis, some modifications have been made to

the datasets. During the merger of the three datasets, the number of observations went down

to 1774 observations due to several reasons. Firstly, the regions (NUTS-2 level) with no

specific regional data (GDP, employment, etc.) were deleted from the database, because for

those regions it was not possible to control for place specific effects. Secondly, the database

has been collapsed based on the NUTS-2 regions, to show the amount of investments per

region (dependent variable) and different business activities. In this way detailed information,

such as company name, of no additional contribution to the model has been deleted. In the

obtained dataset of greenfield investments within European regions, there were regions with

zero greenfield investments and, moreover, for some of the regions no regional data was

available, resulting in exclusion of such data.

Dependent variable

For the empirical analysis, multiple dependent variables will be used. The first one is the

number of greenfield investments per region (NUTS-2) per year, which has been called

‘NumberofInvestments’. The second dependent variable is the logarithm of the ‘Investment

total’, which is the total amount of investments in a certain region in a year. Besides these

dependent variables, attention is given to the specific effect on different business activities

within the relationship between the presence of ports and greenfield investments. To analyse

this effect, several other dependent variables have been made to categorize different

business activities, namely “Business Services”, “Headquarters”, “Logistics”, “Production”,

“R&D”, “Sales & Marketing” and “Support & Servicing”. They indicate the amount of

greenfield investments per year for specific business activities within a region.

Independent variable

To show the effect of ports in a region on the greenfield investments, a dummy variable

‘PortinRegion’ has been introduced based on the freight data of Eurostat. If there was freight

transport in a certain region, the dummy will get a 1 and otherwise a 0.

Control variables

To control for factors that may influence the relationship between the existence of a port in

a region and greenfield investments, we use multiple control variables. In the article of

Karreman et al. (2017), three categories of control variables have been used, namely demand

factors, supply factors and external economies. These categories show the attractiveness of

European regions. In this paper, the first two categories have been used. External economies

category has been left out from the model since there is no accessible data (Table 2). The

control variables are measured on the NUTS-2 regional level and the corporate tax variable –

on the country level. In the correlation matrix (Table 1), where possible multicollinearity has

been checked for.

Table 1 – Correlation Matrix

Demand factors

The demand factors look at market-seeking incentives for companies with possible greenfield

investments. These demand factors consist of GDP per capita, accessibility by air and rail/road

and world cities. GDP per capita is included in the model, because it is expected that this has

positive influence on the relationship between ports and FDI, and companies will more easily

invest in regions which are doing economically well. The accessibility factors are coming from

the ESPON research by Spiekermann and Wegener (2006). Accessibility is important from the

fact that the attractiveness of a region increases when the region is better accessible. The

world city control variable indicates specific investments in a year by region, what type of

world city is present in that region. Introduction of this control variable to the model, enables

to control for the effect that a company will establish itself in a region not only for a port, but

because of a large city in the region which could have multiple benefits, like good company

climate. The world city variable has been divided into 5 groups, namely Alpha+, Alpha, Beta,

Gamma and Rest.

Supply factors

The category supply factor consists of the long-term unemployment, university degree and

corporate tax rate. The long-term unemployment and university degree (education) show the

viability of the labour market (Head and Mayer, 2004). The employment compensation

control variable has not been used, for the reason of high correlation with the GDP per capita

factor. The unemployment factor is coming from Eurostat dataset with NUTS-2 level data. The

university degree factor is measured following the International Standard Classification of

Education (5–6) [ISCED]. It is expected that both these variables will influence the effect of

Port in Region GDP per capita

World city

reference

Long-term

unemployment
University

Accessibility by

air

Accessibility by

rail and road

Congestion costs Corporate tax

Port in Region 1

GDP per capita 0.103 1

World city

reference
-0.010 0.407 1

Long-term

unemployment
-0.151 -0.439 -0.065 1

University 0.153 0.566 0.346 -0.205 1

Accessibility by

air
-0.134 0.574 0.549 -0.171 0.378 1

Accessibility by

rail and road
-0.379 0.314 0.104 0.024 0.091 0.603 1

Congestion

costs
0.057 0.557 0.396 -0.271 0.485 0.57 0.209 1

Corporate tax 0.044 0.284 -0.043 0.048 0.106 0.245 0.45 0.053 1

ports on FDI in a positive way. With long-term unemployment, there are more people in the

labour market available for jobs and with more skilled people (University degree), they will

attract FDI to the region. The corporate tax rate is a good addition to the model, because if

countries have different tax rates, maybe companies will invest in the country with lower

taxes. The corporate tax rate data is deducted from the EY International Tax database

(Brienen et al., 2010). The congestion costs are added as a control variable to the model,

because it could have a negative effect on the relationship of ports on FDI. Congestion cos ts

variable is derived from the ESPON research by Spiekermann and Wegener (2006).

Table 2 – Descriptive Statistics of the Variables in the Models

3.2. Methodology

To examine the effect of the presence of a port(s) in a region on the number of greenfield

investments in that region an appropriate model must be chosen. The dependent variable is

a count variable, thus an OLS model will not be adequate, because count data is discrete and

will normally violate one of the assumptions of OLS, namely the assumption of equal variances

(Gardner et al., 1995). The mean and the variance of count data are linked to each othe r, as

Name Description Mean SD

Port in Region
Dummy variable if there is maritime

(freight) transport in the region with ports
0.38 0.49

GDP per capita Regional GDP (in millions of euros) 22292.6 11047.5

World city

classification

Alpha+ (5), Alpha (4), Beta (3), Gamma (2)

and Rest (1)
1.43 0.93

Congestion

costs
Level of congestion 0.22 0.22

Long-term

unemployment
Long-term unemployment rate 3.18 2.75

University
Percentage of the workforce between

and 64 with tertiary (ISCED 5–6)
0.28 0.09

Corporate taxes Statutory corporate tax rate 29.16 6.08

Accessibility by
air

Number of people that can potentially be

accessed by air (in millions)
96.81 31.97

Accessibility by
rail and road
Number of people that can potentially be

accessed by rail and road (in millions)
98.1 61.22

higher outcomes lead to higher variances. When working with count data, in this case the

number of greenfield investments made in the region, there are mainly two models

appropriate: a Poisson model or a negative binomial model. An important assumption of the

Poisson regression is that the mean is equal to the variance of the count data. However, in

our data this is not the case. The variance is much larger than the mean of the number of

greenfield investment in the regions and therefore this assumption is violated (see table

below). This phenomenon is called overdispersion. Therefore, the Poisson model is not the

right model in this study. In case of overdispersion, the negative binomial model is preferred

and thus this model will be used to examine the effect of the presence of a port in a region

on the number of greenfield investments in that region.

Table 3 – Total number of investments

Variable Observations Mean Standard Deviation

Total number of investments 1891 13.939 23.076

Another common problem with count data is that the data is skewed due to the presence of

many outliers in the data. This could occur due to the presence of multiple high values, but

most of the time skewness with count data is caused by many zeros in the data (Atkins et al.,

2013). In the latter case a zero-inflated negative binomial model must be chosen. However,

in the data obtained in this study there are no values of zero. Consequently, this leads to a

new situation in which the model must correct for the fact that there are no zeros in the count

variable. This is possible using the zero-truncated model, which is used when count data is

used and a value of zero is not present in the dataset. Keeping in mind that overdispersion is

present in the count data, the definitive model is a zero-truncated negative binomial.

TotalNumberofInvestmentsit = β0 + β1PortinRegionit + β2GDPpercapit +

β8LongtermUnemploymentit + β3Universityit + β5AccessibilitybyAirit + β6AccessibilitybyRailRoadit

+ β7CongestionCostsit + β8CorporateTaxit + β4WorldCityReferenceit + εit

Also, in this study a model is used to estimate the effect of the presence of a port on the total

amount of greenfield investments in that region. Since the total amount of investments is a

continuous variable a negative binomial model would not be logical, instead a fixed effects or

random effects model is more appropriate. The fixed effects model will omit all the time –

invariant variables, which will not be beneficial for the model (Allison, 2009). For example,

both infrastructure variables will be dropped as well as the world city classification. More

importantly, the main variable of interest, namely the presence of a port in the region, will be

omitted in this model, because of its time-invariant character. The fixed effects model will

therefore not be the right model to use in the present study. Based on this large disadvantage,

the random effects model is chosen to estimate the effect of the presence of a port on the

total amount of greenfield investments in that region. It is expected that this estimation yields

the same sign and significance for the main variable, namely the presence of the port. In both

above mentioned models several control variables will be included.

Thereafter, three models will be run to test hypothesis two and three. All these random

effects model will have an interaction variable, which will indicate if the hypothesis two and

three hold. In the first of these models an interaction variable is created with the presence of

the port in a region and the world city classification to test hypothesis two. In the latter two

of these models an interaction variable is created with the presence of the port in a region

multiplied by on the one hand accessibility by air and on the other accessibility by rail and

road. Hence, the outcome will show if the hypothesis for infrastructure holds. Furthermore,

for all random effects models the standard errors are made robust to enhance the structural

validity.

The effect of the presence of a port in a region on number of greenfield investments for

different business activities will also be researched. Seven random effects regressions will be

performed in which the different types of business activities will be used as dependent

variable. Following this methodology will lead to the results and hence helps to answer our

hypothesis and consequently the main question.

4. Results

In this section, the results of the truncated negative binomial model concerning the effect of

the presence of a port in a region on the number of greenfield investment in that same region

will be discussed firstly. Secondly, the random effects model regarding the effect of the

presence of a port in a region on the total amount of investments in that same region will be

addressed. In the random effects models, multiple interaction effects are used to check

whether hypothesis 2 and 3 hold. To examine the effect of different business activities on the

presence of ports in a region on FDI, multiple random effects model will be used. This model

will conclude the section and afterwards the results will be discussed in the discussion.

Table 4 shows the results of the truncated negative binomial model and two random effects

model with the world city interaction effect. In the first model, the dependent variable is the

TotalNumberofInvestments. The independent variable, which indicates if there is a port in the

region, is presented as a dummy variable and takes the value of ‘1’ if a port in present in the

region and a value of ‘0’ if this is not the case. The results show that the presence of a port in

a region positively affects the number of greenfield investments in that same region. This

finding is significant at a 1% significance level and means that the logs of expected counts

would increase with 0.18, ceteris paribus. This finding supports our first hypothesis.

Moreover, almost all control variables are significant at a significance level of 1% in this first

estimation, only long-term unemployment is not significant.

However, there are some unexpected signs regarding the control variables. For example, the

variable GDP per capita has a negative sign, where a positive sign was expected. When the

GDP per capita increases in a region, it has been expected that this would attract greenfield

investments due the fact that there will be more demand in the region. However, the results

argue otherwise. Also, the variable congestion costs show a peculiar effect. The coefficient

shows that an increase in congestion costs will lead to an increase in the number of greenfield

investments. This is peculiar, as when reasoned logically, higher costs usually make a region

less attractive for investments. Again, the results provide evidence in the contrary. The

coefficients of other variables are also notable given their large positive coefficients, namely

the dummy variable of the world city classifications. In this model, the effect of an Alpha+ city

classification leads to an increase in the logs of expected counts of 2.86, ceteris paribus.

The results of the random effects model (model 2) which estimates the effect of the presence

of a port in a region on the total amount of greenfield investments in that same region.

Correspondingly with the results of the truncated negative binomial model in Table 4, the

random effects model (model 2) also shows a positive and significant effect of the variable

port in the region.

Table 4 – Results TNB Binomial model & random effects models

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table 5 – Random effects models by accessibility

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

The GDP per capita variable shows the expected effect in the random effects model, namely

a positive effect on the total amount of investments. However, long-term unemployment and

corporate tax show an opposite effect compared to the negative binomial model. This is also

contradictory with our expectation. The rest of the variables are compliant with our

expectations, such as congestion costs (negative effect).

The third model is included with the interaction effect between port(s) in a region and

different levels of world cities. All the interaction effects for every level of world city are

insignificant. Comparing model 3 with model 2, hardly any changes can be discovered and

eventually hypothesis 2 can be rejected.

Table 5 includes the fourth model, where an interaction effect with port in the region and

accessibility by air is added; and the fifth model, where an interaction effect with port in a

region and accessibility by rail & road is included. All these interaction effects are insignificant.

Comparing model 4 and 5 with model 2, hardly any changes can be discovered and eventually

hypothesis 3 must be rejected.

Table 7 shows the results of the seven random effects models, with the different business

activities as dependent variable. The estimates of the effect of the presence of a port in a

region are of main interest, since this outcome will show if hypothesis 4 holds. The results for

the variable presence of port are mostly insignificant, however for the logistics sector the

variable is positive and significant at a significance level of 1%. So, when a port is present in a

region an increase in logistics greenfield investments is expected. This partially aligns with the

expectations mentioned in the literature review. Table 6 is a presentation of the shares of

business activities in the overall amount of investments. It shows the distribution of

investments in our dataset.

Table 6 – Function distribution of greenfield investments in European regions

Table 7 – Random effects models by business activities

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

The control variables show some different effect, then in the first random effects model. GDP

per capita shows mixed results as well as congestion costs. However, other variables show

predominantly the same sign and significance. For example, the education variable,

university, is positive and significant for six of the seven models. The same goes for the world

city classifications variables. All coefficients are positive and only 3 out of 28 coefficients are

not significant. The corporate tax level is negatively significant for all models. Long-term

unemployment, accessibility by air and rail and road are predominantly insignificant.

5. Conclusion and Discussion

When firms want to invest in other countries and/or regions, they look at several

determinants of location choices. A possible factor in these location choices is the presence

of a port. In this paper, we present evidence for the relationship between the presence of a

port and foreign direct investments in European region. In detail, the main research question

namely whether and to what the presence of a port affects the greenfield FDI in the European

NUTS-2 regions, has been studied from four angles. The first angle, is the positive presence of

a port in a region and foreign direct investments. Second, we address the effect of a world

city in the region with the presence of a port and relation to FDI. Third, we check whether the

accessibility by air and rail & road have a positive influence of the relationship. Finally, we

made a distinction between multiple business activities of the investments and the presence

of a port.

The main results of the empirical analysis have a mixed outcome. The presence of a port in a

region has in model 1 till 3 a positive significant influence on investments in that same region,

but for models 4 and 5 there is no significant evidence on this effect. Focussing more on detail

with also the presence of a world city and accessibility by air and rail & road in the region, we

don’t find evidence that these factors have an influence on the relationship between ports

and FDI. Looking at the differences in business activities and investments, we find only a

positive outcome for logistics. This is an expected outcome, following the fact that ports are

important nodes in the supply chain of many firms (Pettit and Beresford, 2009).

Concerning the control variables some outcomes were contradictory to the expectation. For

example, the GDP per capita in the negative binomial model is negatively significant, where

the expectation would be that higher a higher GDP per capita would lead to more greenfield

investments in that same region. Since the average employee compensation in the dataset is

highly correlated with the GDP per capita, a possible explanation could be that higher wages

deter greenfield investments. Another unexpected result in the negative binomial model was

the positive and significant coefficient for the congestion costs. Logically, higher costs lead to

a lower attractiveness of a region, however maybe due to the popularity of region, these costs

are taken for granted. In the random effects model, other remarkable results were obtained.

The long-term unemployment variable was negative and significant. Higher unemployment

means a wide offer of labour forces. However, this could also be a sign that the region is not

performing that well. Another peculiar effect is that the corporate tax rate variable shows a

positively significant effect. Lower taxes usually lead to a higher attractiveness of a region. A

possible explanation could be that higher corporate tax rates are correlated with better legal

systems, moreover the infrastructure was also somewhat correlated with the level of

corporate taxes. These results can serve as the basis for future research.

The overall answer to the research question is that there is no or little evidence in this study

that the presence of a port in a European region lead to more greenfield investments in the

same European region. None of the hypothesis can be fully accepted and some even need to

be rejected. A limitation in this research is regarding the size of the port in a region. In a

future research, it would be good if an independent variable would be created with different

levels of the size of ports and their effect on greenfield investments in the same region.

Another limitation is that there are no control variables in the field of external economies, in

which we control for previous investments in the same region. Also, concerning previous

investments in specific business activities, such as logistics. Future research could be focussed

on other continents, since this research has only included European regions.

6. References

 Allison, P. D. (2009). Fixed effects regression models (Vol. 160). SAGE publications.

 Aschauer, D.A. (1989). Is public expenditure productive? J. Monetary Econ. 23, 177–200.

 Atkins, D.C., Baldwin, S.A., Zheng, C., Gallop, R.J., & Neighbors, C. (2013). A tutorial on

count regression and zero-altered count models for longitudinal substance use data.

Psychol Addict Behav. 2013 Mar; 27(1): 166-77. doi: 10.1037/a0029508. Epub 2012 Aug

20.

 Bagchi-Sen, S. (1997). Service employment in large, medium, and small metropolitan areas

in the United States. Urban Geography, 18(3), 264-281.

 Ballou, R. H. (2007). The evolution and future of logistics and supply chain management.

European business review, 19(4), 332-348.

 Banuri, T., Schor, J. (eds.) (1992). Financial openness and national autonomy:

Opportunities and constraints. Oxford. Clarendon Press.

 Basile, R., Castellani, D., & Zanfei, A. (2009). National boundaries and the location of

multinational firms in Europe. Papers in Regional Science, 88: 733–748.

 Beaverstock, J. V.& & Smith, J. (1996). Lending jobs to global cities: skilled international

labour migration, investment banking and the City of London. Urban studies, 33(8), 1377-

1394.

 Beaverstock, J., Smith, R., & Taylor, P. (1999). A roster of world cities. Cities, 16(6): 445-

458.

 Boopen, S. (2006). Transport infrastructure and economic growth: evidence from Africa

using dynamic panel estimates. The empirical economics letters, 5(1), 37-52.

 Bottasso, A., Conti, M., Ferrari, C., & Tei, A. (2014). Ports and regional development: a

spatial analysis on a panel of European regions. Transportation Research Part A: Policy

and Practice, 65, 44-55.

 Brienen, M. J., Burger, M. J., & van Oort, F. G. (2010). The Geography of Chinese and Indian

Greenfield Investments in Europe. Eurasian Geography and Economics, 51(2).

 Burger, M. J., van der Knaap, B., & Wall, R. S. (2012). Revealed competition for greenfield

investments between European regions. Journal of economic geography, 13(4), 619-648.

 Budd, L. (1998). Territorial competition and globalization: Scylla and Charybdis of

European cities. Urban Studies, 35: 663–685.

 Cheshire, P. C., & Gordon, I. R. (1995). European integration: the logic of territorial

competition and Europe’s urban system. In J. Brotchie, M. Batty, E. Blakely, P. Hall, P.

Newton (eds) Cities in Competition: Productive and Sustainable Cities for the 21st Century,

pp. 108–126. Melbourne: Longman.

 Cheshire, P. C., & Gordon, I. R. (1998). Territorial competition: some lessons for policy. The

Annals of Regional Science, 32: 321–346.

 Cho, H., & Ha, Y. (2009). Determinants of FDI Inflow in Regional Port with Resource-Based

View and Institutional Theory: A Case of Pohang-Yeongil Port. The Asian Journal of

Shipping and Logistics, 25(2), 305-331.

https://www.ncbi.nlm.nih.gov/pubmed/22905895

 Condeco-Melhorado, A., Gutierrez, J., Garcia-Palomares, J.C. (2011). Spatial impacts of

road pricing: accessibility, regional spillovers and territorial cohesion. Transportation

Research Part A 45, 185–203.

 Démurger, S. (2001). Infrastructure development and economic growth: an explanation

for regional disparities in China?. Journal of Comparative economics, 29(1), 95-117.

 Ducruet, C. (2006). Port-city relationships in Europe and Asia. Journal of International

logistics and trade, 4(2), 13-35.

 Dunning, J. H. (1993). Multinational enterprises in the global economy. Wokingham, UK:

Addison-Wesley.

 Dunning, J. H. (1998). Location and the multinational enterprise: a neglected factor?

Journal of International Business Studies, 29: 45–66.

 European Spatial Planning Observation Network – ESPON, 2009. Accessibility Project

Reports. .

 Fujita, M., Mori, T. (1996). The role of ports in the making of major cities: self-

agglomeration and hub-effect. J. Dev. Econ. 49 (1):93–120.

http://dx.doi.org/10.1016/0304-3878(95)00054-2.

 Gardner, W., Mulvey, E.P., & Shaw, E.C. (1995). Regression analyses of counts and rates:

Poisson, overdispersed Poisson, and negative binomial models. Psychol Bull. 1995 Nov;

118(3): 392-404.

 Goerzen, A., Asmussen, C. G., & Nielsen, B.B. (2013). Global cities and multinational

enterprise location strategy. Journal of International Business Studies, 44 (5), The

Multinational in Geographic Space (June/July 2013), pp. 427-450.

 Gordon, I. R. (1999). Internationalisation and urban competition. Urban Studies, 36: 1001–

1016.

 Hall, P. V. (2007). Seaports, urban sustainability, and paradigm shift. Journal of Urban

Technology, 14(2), 87-101.

 Haralambides, H.E. (2012). Ports: Engines for Growth and Employment. European Ports

Policy Review: Unlocking the Growth Potential. European Commission, Brussels, 25 -26

September 2012.

 Head, K., & Mayer, T. (2004). Market potential and the location of Japanese investment in

the European Union. Review of Economics and Statistics 86 (4), 959 – 972.

 Jacobs, W., Ducruet, C., & De Langen, P. (2010). Integrating world cities into production

networks: the case of port cities. Global networks, 10(1), 92-113.

 Jungnickel, R. (1993). Recent trends in foreign direct investment, in Bailey, P., Parisotto, A.,

Renshaw, G. (eds.) (1993). Multinationals and employment: The global economy of the

1990s. Geneva. ILO.

 Karreman, B., Burger, M. J., & van Oort, F. G. (2017). Location Choices of Chinese

Multinationals in Europe: The Role of Overseas Communities, Economic Geography, 93:2,

131-161, DOI: 10.1080/00130095.2016.1248939

 Krugman, P. (1991a). Geography and trade. Cambridge, MA: MIT Press.

https://www.ncbi.nlm.nih.gov/pubmed/7501743

 Krugman, P. (1991b). Increasing returns and economic geography. J. Polit. Econ. 99, 483–

499.

 Krugman, P. (1993). The hub effect: or, threeness in international trade. In: Ethier, W.,

Helpman, E., Neary, P. (Eds.), Theory, Policy and Dynamics in International Trade.

Cambridge University Press, Cambridge.

 Lever, W. F. (1999). Competitive cities in Europe. Urban Studies, 36: 935–948.

 Li, S., Ragu-Nathan, B., Ragu-Nathan, T. S., & Rao, S. S. (2006). The impact of supply chain

management practices on competitive advantage and organizational performance.

Omega, 34(2), 107-124.

 Maggi, E., & Mariotti, I. (2012). Globalisation and the rise of logistics FDI: the case of Italy.

Foreign Investment: Types, Methods and Impacts, 29-60.

 Markusen, A. (1996). Sticky places in slippery space: A typology of industrial districts.

Economic Geography, 72(3): 293–313.

 Markusen, A. R., & Nesse, K. (2007). Institutional and political determinants of incentive

competition. In A. R. Markusen (ed.) Reining in the Competition for Capital, pp. 1–42.

Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.

 Martin, C. (1992). Logistics and supply chain management, (pp. 4). Financial Times,

London.

 Musso, E. (2004). Porti, lavoro, economia: le regioni portuali di fronte alla rivoluzione

logistica. CEDAM.

 Nachum, L., & Wymbs, C. (2005). Product differentiation, external economies and MNE

location choices: M&A in global cities. Journal of International Business Studies, 36(4):

415–434.

 Niebuhr, A. (2006). Market access and regional disparities: new economic geography in

Europe. Ann. Regional Sci. 40, 313–334.

 Notteboom, T. E., & Rodrigue, J. P. (2005). Port regionalization: towards a new phase in

port development. Maritime Policy & Management, 32(3), 297-313.

 O’Connor, K. (1989). Australian ports, metropolitan areas and trade-related services. The

Australian Geographer, 20(2), 167-72.

 OECD, (2014). The Competitiveness of Global Port-cities. OECD Publishing, Paris

http://dx.doi.org/10.1787/9789264205277-en.

 Ottaviano, G. (2008). Infrastructure and economic geography: an overview of theory and

evidence. Eur. Bank Invest. Pap. 13, 8–35.

 Pettit, S. J., & Beresford, A. K. C. (2009). Port development: from gateways to logistics hubs.

Maritime Policy & Management, 36(3), 253-267.

 Porter, M. (2001). Regions and the new economics of competition. In A. Scott (Ed), Global

city-regions: Trends, theory, policy: 139–152. Oxford: Oxford

University Press.

 Puga, D. (2002). European regional policies in light of recent location theories. J. Econ.

Geogr. 2, 373–406.

 Raines, P. (2003) Flows and territories: the new geography of competition for mobile

investment in Europe. In N. Phelps, P. Raines (eds) The New Competition for Inward

Investment, pp. 119–136. Cheltenham: Edward Elgar.

 Rietveld, P. (1994). Spatial economic impacts of transport infrastructure supply.

Transportation Research Part A: Policy and Practice, Elsevier, vol. 28(4), pages 329-341,

July.

 Robinson, R. (2002). Ports as elements in value – driven chain systems: The new paradigm.

Maritime Policy and Management, 19: 3–8.

 Rodrigue, J. P., Comtois, C., & Slack, B. (2013). The Geography of Transport Systems (pp.

142), Hofstra University, Department of Global Studies & Geography.

 Sassen, S. (1991). The global city: New York, London, Tokyo. Princeton, NJ: Princeton

University Press.

 Sassen, S. (2000). The global city: strategic site/new frontier. American studies, 41(2/3),

79-95.

 Sheffi, Y. (1985). Urban transportation networks. Englewood Cliffs, NJ: Prentice Hall.

 Slack, B. (1989). Port services, ports and the urban hierarchy. Tijdschrift voor Economische

en Sociale Geografie, 80(4), 236-243.

 Spiekermann, K., & Wegener, M. (2006). Accessibility and spatial development in Europe.

Scienze Regionali 5 (2), 15 – 46.

 Turok, I. (2004). Cities, regions and competitiveness. Regional Studies, 38: 1069–1083.

 UNCTAD, (1993). World Investment Report 1993: Transnational corporations and

integrated international production. New York. United Nations.

 Zhao, Q., Xu, H., Wall, R. S., & Stavropoulos, S. (2017). Building a bridge between port and

city: Improving the urban competitiveness of port cities. Journal of Transport Geography,

59, 120-133.

Turn in your highest-quality paper
Get a qualified writer to help you with

“ FDI & Ports ”

Get high-quality paper

NEW! AI matching with writer

Order an Essay Now & Get These Features For Free:

Turnitin Report

Formatting

Title Page

Citation

Outline

Place an Order