You are on page 1of 26

European Planning Studies, 2014

Vol. 22, No. 2, 268 292, http://dx.doi.org/10.1080/09654313.2012.731038

Factors Influencing Location Selection of


Warehouses at the Intra-Urban Level:
Istanbul Case
ALI DURMUS , & SEVKIYE SENCE TURK

Sinpas GYO, Istanbul, Turkey, Department of Urban and Regional Planning, Faculty of Architecture,
Istanbul Technical University, Istanbul, Turkey

(Received June 2011; accepted September 2012)

ABSTRACT Depending on the worldwide developments in the industry and services sectors, the
increased significance of the logistic sector is continuing. In parallel to this tendency, the need
for warehousing as one of the basic activities of logistics has increased and the warehouses have
returned to the important function areas for urban sites. In recent years, warehouses are
demanded in the real estate market just like houses and offices. In the literature, although there
are many supply chain studies focused on location selection of warehouses, there is a lack of
studies discussing the behaviour of warehouse location at the intra-urban level. The article
investigates the factors affecting the location selection of warehouses and their degree of
importance in Istanbul as a case area by using a logistic regression model. It was found that
location-specific factors are effective on the location selection of warehouses in the Istanbul
metropolitan area, and warehouse location follows a certain economic rationality at the intraurban level.

Introduction
Logistics refers to the process of managing the flow of materials from point of origin to
point of consumption (Hong, 2007). The logistics plays a vital role for the global
economy. Global logistics expenditures represented 13.8% of the worlds GDP for the
year of 2002 (Rodrigues et al., 2005). Also, the global logistics market generated total revenues of $3566 billion in 2008, representing a compound annual growth rate of 6% for the
period spanning 2004 2008. The sector is forecast to grow at a rate of 2% for the 5-year
period 2008 2013, to reach $3895.5 billion by the end of 2013 (Datamonitor, 2009).

Correspondence Address: Sevkiye Sence Turk, Department of Urban and Regional Planning, Faculty of Architecture, Istanbul Technical University, 34437 Tasksla Taksim, Istanbul, Turkey. Email: senceturk@gmail.com

Current Address: Ali Durmus, Middle Anatolia Devlopment Agency, Kayseri, Turkey.
# 2012 Taylor & Francis

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

269

Warehousing is one of the most important activities in logistic services. Warehouses are
logistic units where commodity surplus is stored for short or long terms under sanitary
conditions, and various finished and semi-finished products are stored for distribution
before production. Also recycling materials are classified or prepared for destruction,
and goods of trade companies are stored before transportation (Demirel et al., 2010).
Warehouse location is one of the most important and strategic decisions in the optimization of logistic systems. For example, while freight transport is only around 40% of
the logistic costs of firms, most of the remaining logistics expenditures go for storage,
correct handling and the financing of inventory (McKinnon, 2009). The location selection
of warehouses directly affects the form and structure of the whole logistics system of a
firm. In the literature, while a lot of studies focus on location selection of warehouses
based on multi-criteria decision making among alternative locations (Eilon, 1982;
Korpela & Touominen, 1996; Korpela et al., 2007; Drezner et al., 2003; Melo et al.,
2009; Demirel et al., 2010), a limited number of studies focus on the distribution of warehouses at the national level or intra-metropolitan level (Glasmeier & Kibler, 1996, p. 740;
Sivitanidou, 1996; Bowen, 2008; Cidell, 2010).
Warehousing establishments are unsurprisingly concentrated in metropolitan areas.
Warehouses are now demanded on the real estate market like houses and offices. In
this sense, location selection of warehouses can affect the structure and form of the
city. However, how logistic warehouses are distributed or what their logic for location
is at the intra-urban level is not exactly known. Empirical studies on location selection
are limited at the intra- urban level. The aim of this article is to examine the behaviour
of location selection in the logistic sector at the intra-urban level, focusing on Istanbul
as a case area. In the article, factors affecting selection and their degree of importance
are determined by using a logistic regression model. The impact of warehousing on
the urban area is discussed. The findings are expected to contribute to the decision
makers for guidance in land use plan decisions and to investors for investment
decisions.
Developments in world trade and increased foreign trade with Turkeys export-based
growth strategy after the 1980s have led to important developments in the Turkish logistic sector (IMP, 2006a; Ozdemir Darby, 2010). Despite the corporate structure and
BITAK, 2006), the logistic sector
deficiencies in the infrastructure (Gurdal, 2006; TU
in Turkey has reached the capacity in supplying services to actual and potential
markets in its region (Babacan, 2005; Ersoy, 2006; Ozdemir Darby, 2008, 2010). The
increased importance of logistic services has made this sector attractive to entrepreneurs
in terms of investments. Domestic firms have taken important steps to supply logistic
services. International firms have entered the Turkish market and started taking part
in the sector through acquisitions, mergers or direct capital investments (IMP, 2006a,
2006b, 2006c; IMM, 2008). The logistic sector has an important share (14%) of the
GNP in Turkey. There has been a considerable increase in the share in the GDP of
the logistic sector for a decade. The size of the sector which was 46,490,154 TL,
reached 1,654,588,618 TL in 2005 (Turkstat, 2010). Istanbul holds an important place
in the logistics sector in Turkey as 60% of all logistic activities are realized in Istanbul
(IMP, 2006a, 2006b, 2006c). In recent years, because of increasing demand, the logistic
warehouses have begun to be used in the real estate sector as a tool of investment
(Colliers International, 2010; Durmus, 2010). These characteristics can make Istanbul

270

A. Durmus & S.S. Turk

a good laboratory for examining the distribution of warehouses within an urban area and
the warehouse location.
In the second section of the article, following the introduction, the literature on
factors affecting location selection is analysed. The third section reviews the development of logistic warehouses and their spatial distribution in Istanbul. The fourth
section includes a modelling for location selection of warehouses in Istanbul. This
section is divided into five sub-sections. The sub-sections give, respectively, the methodology, hypotheses and explanatory variables, data, model specification and the findings obtained from the model. The fifth section of the article is reserved for general
evaluation and conclusions.

Literature Review
Main Approaches of Location Selection of Warehouses
In the literature, the previous studies related to location selection of warehouses have been
addressed in two different ways. First are the studies that examine the distribution of the
warehouse at the national level or intra-metropolitan level. These studies mostly focused
on the changing geography of warehousing. However, these are quite limited (Glasmeier
& Kibler, 1996, p. 740; Sivitanidou, 1996; Bowen, 2008; Cidell, 2010). For example,
Bowen (2008) studied to discern the degree to which the expansion of warehousing has
gravitated towards places with superior accessibility in the nations air, maritime, rail
and highway transportation networks. Cidell (2010) also examined the sub-urbanization
of warehousing and trucking activity within US metropolitan areas between the 1980s
and the present using Gini indices as a measure of concentration. Sivitanidou (1996) examined whether the rent for warehouse space for locations across the region depends on a
handful of factors such as access to an airport, density of the local freeway network and
proximity to a major freeway junction.
Second are the studies that examine the location selection of warehouses with a multicriteria decision-making methodology, choosing the decision making of firms best warehouse location among many alternatives. Multi-criteria decision making can be defined as
the evaluation of the alternatives for the purpose of selecting or ranking, using a number of
qualitative and/or quantitative criteria that have different measurement units (Ozcan et al.,
2011). The location selection of warehouses is evaluated with the different analysis methodologies such as AHP, TOPSIS, ELECTRE I, ELECTRE II and ELECTRE III (Korpela
& Touominen, 1996; Drezner et al., 2003; Demirel et al., 2010; Ozcan et al., 2011). In
each analysis, the methodology is different. For example, while AHP based its methodology on a small number of alternatives and criteria, quantitative or qualitative data,
ELECTRE III includes objective and quantitative data, and usage of fuzzy logic (Kahraman et al., 2003; Tabari et al., 2008). Vlachopoulou et al. (2001) underline the significance
of a geographic decision support system for the warehouse site selection process in order
to enable the manager to use quantitative and qualitative criteria to classify alternative
warehouses.
The article is based on modelling of distribution of warehouses at the intra-metropolitan
level instead of focusing on best warehouse location among many alternatives with
multi-criteria decision-making methodology.

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

271

Main Criteria Used in the Location Selection of Warehouses


In the studies focused on the distribution of the warehouse at the national level or the intrametropolitan level, generally, most of the factors in the location of warehousing activity
are similar to any other industry. These are: proximity to customers or clients, reasonable
real estate costs, access to interstate highways, availability of appropriately skilled
workers and reasonable costs of doing business (Glasmeier & Kibler, 1996, p. 740). At
the national level within the USA, close proximity to the Central Business District
(CBD), to other warehouses and industrial facilities and to the pre-interstate transportation
routes of rivers and rail was found to be the most important factor (Eaton, 1982). Also,
Bowen (2008) found that access to air, highways and rail led to more warehousing establishments in a county in 1998 while only air and highway access were significant in 2005.
Access to rail and port facilities were found not significant.
At the metropolitan level, the study by Sivitanidou (1996) of warehouse and distribution
establishments within greater Los Angeles found higher rents for properties that were
larger, in closer proximity to market and production locations, had better highway and
airport access (but not better port or rail access) and were closer to blue-collar workers.
Similarly, the model of Woudsma et al. (2008) of logistics-related land uses in Calgary
found that access to highways and the airport led to more logistics development while
access to the central city was irrelevant.
In the studies focused on the location selection of warehouse with multi-criteria
decision-making methodology, criteria as costs, labour characteristics, infrastructure,
markets and macro-environment were used (Alberto, 2000; MacCarty & Atthrirawong,
2003; Demirel et al., 2010). The study by Vlachopoulou et al. (2001) used different criteria
for location selection of warehouses. These are the actual and potential customers in the
surroundings (size of the market), the purchasing power of this market (GDP), means of
access to the region (accessibility options), other competitors (agglomeration) in the
region and the infrastructure and superstructure capacity.
In the article, transportation accessibility, market size, clusters, distance to the city
centre, rents and customs are used as criteria. Commonly, accessibility is one of the important criteria. The location selection of warehouses providing the best use of roads, railways, air and maritime transport can be the optimum solution in terms of cost, time and
customer satisfaction. Accessibility plays an important role in minimizing the transport
costs (Dogan, 2006; Demirel et al., 2010).
Also, one of the important criteria is market size. Generally, GDP is a variable that
shows the economic output of the urban district. For certain investments, the GDP
plays a determinant role (Wu & Radbone, 2005). The high level of development and
welfare in districts can imply that service and industry sectors are more active. In developed districts, as commercial activities and consequently load circulation are high, concentration of warehouses is expected to increase.
Clusters are one of the important criteria. Logistics and storage cannot be in a place
where there is no production. When products collected from the factory in order to be
delivered to the end-user or the agent are gathered in the nearest warehouse, transport
costs can be further reduced. Similarly, a warehouse established in an optimal location
between raw materials and production facilities can decrease costs. In this sense, increase
in the concentration of warehouses in regions of industrial clusters can be expected
(Fulton, 1971). However, the same situation cannot be valid for clusters in the services

272

A. Durmus & S.S. Turk

sector. In general, since rentals are high in regions where FIRE (Finance, Insurance and
Real Estate) activities are effective, they are not preferred for warehouses.
Distance to the city centre can be considered as one of the criteria. CBDs are important
focal points that maximize the possibility of location advantages such as proximity to
wider markets, proximity of other business services, access to educated workforce,
clients and suppliers and possibilities for collaboration, competition and information
exchange (Bodenam, 1998; Shearmur & Alvergne, 2002; Polese & Shearmur, 2004).
The CBD appears to be the most highly specialized centre of high-order services
(Coffey et al., 1996, Shearmur & Coffey, 2002). In a market economy, the distance to
the CBD is usually an important determinant of residential and industrial location. Land
prices decline away from the CBD. However, in the polycentric urban areas, a landprice profile cannot be explained by the distance to the CBD (Heikkila et al., 1989;
Wu, 2000). Due to disadvantages such as traffic density, high rents and being far from
industrial clusters, CBDs can be considered not preferable in the choice of warehouse
location.
Rents can also be considered as one of the criteria. Although suitable locations can be
chosen for warehouses in order to reduce transport costs, this advantage cannot be offset
by the high rentals of industrial facilities in the chosen area. Otherwise, a warehouse which
is close to all means of transport and the city centre can increase costs. Also, customs can
be considered as one of the criteria. In order to facilitate entry and exit from the country
and completion of transport procedures, it can be considered that logistics firms choose to
be close to customs.

Significance of Warehouses for Istanbul and Spatial Distribution of Warehouses


Significance of Warehouses for Istanbul
Istanbul is Turkeys trade, finance, tourism and industrial centre and is a regional centre for
the neighbouring countries. According to 2011 data, Istanbul generates 23% of the GDP
(Standard & Poors, 2011). In Istanbul, the distribution of the labour force according to
sectors is as follows: industry 39.9%, services 59.6% and agriculture 0.4% (Turkstat,
2010). According to the data in 2009, 29% of the added value in the industry sector in
Turkey is produced by Istanbul. While, in Turkey, 9.4% of the GDP is from the agriculture
sector, 28.2% of the GDP is from the industry sector and 62.4% of the GDP is from the
service sector, 0.3% of the added value from agriculture, 29.1% of the added value
from industry and 70.6% of the added value from services in Istanbul are produced (Ministry of Information, Industry and Technology, 2009). This demonstrates that Istanbul is an
important centre for services and industry. In Turkey, 60% of logistics operations are realized in Istanbul. All of these show the importance of the logistics sector for Istanbul and
the increased need for warehouses (IMM, 2006). Especially after the 1980s, reasons such
as the development of industry in the city, the increase in production, the increase in its
import and export capacity at the global scale and the necessity to keep goods in stock
at optimum levels in order to increase consumer satisfaction have made warehouses indispensable for the industry and services sectors.
In Istanbul, the share of persons employed in the industrial sector within the active
labour force is 39.9%. In the sub-sector distribution in the industry sector, clothing is
first with 17%, textiles is second with 11% and machines and machine parts are third

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

273

with 11% (Ministry of Information, Industry and Technology, 2009). For this reason, a
great number of the actual warehouses in Istanbul are used for textiles and yarns, machines
and machine parts.
Considering the impact on industrial storage activities and the industrial cluster, the
Organized Industrial Zones in Istanbul have a great share (Ministry of Information, Industry and Technology, 2009) (Figure 1). The greatest organized industrial zone in Istanbul is
the Ikitelli Organized Industrial Zone built on an area of 700 ha. Industrial clusters in the
mraniye on the Asian side, and in the textiles sector in
metal sector are concentrated in U
Kucukcekmece on the European side (IMM, 2006d; Gezici et al., 2008). Therefore, industrial clusters benefiting from the advantage of logistics facilities in those regions affect
favourably the location selection of warehouses.
In Istanbul, the share of the services sector increased in the distribution of the sectoral
distribution of the labour force. Of the total number of people in active employment, some
59.9% work in services, including construction and electricity gas services, besides the
main services sector. This figure increased slightly from 1990 to 2000. If construction
and electricity gas services are excluded, the share of services drops to 51% for 1980,
51% for 1990, 53% for 2000 (IMM, 2008) and 59.6% for 2010 (Turkstat, 2010). According to records of the Istanbul Chamber of Commerce in 2005 and 2006, there are 285,997
firms in total in Istanbul. Of these firms, 15% in producer services, 80% in distributive services, 4% in personal services and 1% in social services are active. The places where the
services sector is concentrated are, respectively, Kadkoy (14%), Sisli (9%), Eminonu
(7%) and Besiktas (5%) (IMM, 2006). Also, service clusters benefiting from the advantage
of logistics facilities in those regions affect favourably the location selection of warehouses.
According to the needs of the industry and services sectors in Istanbul, the use of warehouses shows the change, demonstrating differences from one district to another. The
differences in the districts are seen in Table 1. According to Figure 1, the warehouses
in the districts like Besiktas, Eminonu, Fatih and Bakrkoy that can identified as central

Figure 1. Classification of warehouses according to sub-sectors.

274

A. Durmus & S.S. Turk


Table 1. The number of warehouses according to districts in Istanbul

Percentage

The number of warehouses in the


model (the addresses of which have
been obtained)

Percentage

596
15
90
475
784
51
45
20
2076

7.7
0.1
1.2
6.2
10.1
0.7
0.5
0.2
26.7

0
15
93
0
758
26
42
0
934

0
0.4
2.2
0
17.8
0.6
0.9
0
21.9

Suburban districts
Bahcelievler
K.C
ekmece
Gungoren
Bagclar
B.Cekmece
mraniye
U
Tuzla
Maltepe
Kartal
Kagthane
Zeytinburnu
Bayrampasa
Beykoz
Avclar
Pendik
Saryer
Eyup
Gaziosmanpasa
Catalca
Silivri
Esenler
Sultanbeyli
Sile
Total

242
487
77
345
107
228
196
106
148
193
985
1403
36
25
81
16
95
270
56
266
294
51
n.a.
5707

3.1
6.2
0.9
4.4
1.4
2.9
2.5
1.4
1.9
2.5
12.7
18.1
0.5
0.3
1
0.2
1.2
3.5
0.7
3.4
3.7
0.6
0
73.3

242
541
55
0
170
87
160
0
126
0
0
1584
36
57
0
0
0
0
23
3
186
55
n.a.
3325

5.7
12.7
1.3
0
4.0
2.0
3.8
0
3.0
0
0
37.2
0.8
1.3
0
0
0
0
0.5
0.1
4.4
1.3
0
78.1

General total

7783

Number of
warehouses
Central districts
Sisli
Besiktas
Beyoglu
Kadkoy
Eminonu
Fatih
Bakrkoy
skudar
U
Total

100

4259

100

n.a., Data not available.

districts are used by the services sector. For example, 100%, 95%, 80%, 80% and 65% of
warehouses in Fatih, Eminonu, Besiktas, Bakrkoy and Beyoglu, respectively, are used by
the services sector (Figure 1). In these districts, the number of warehouses used by the
industry sector is quite low. On the other hand, the warehouses in the districts like
Tuzla, Kartal, Kucukcekmece, Gungoren, Avclar, Bahcelievler, Bayrampasa, Buyukcekmece and Silivri that can be identified as suburban districts are generally used by industry
(Figure 1). For example, 95%, 75%, 70% and 70% of warehouses in Tuzla, Kartal, Buyuk-

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

275

cekmece and Gungoren, respectively are used by the industry sector (Figure 1). Differences in the use of warehouses can be explained by dependence on the location of services
and industry sectors in Istanbul. It is known that the services sector is dominant in central
districts. Warehouses in suburban districts are mostly used by the industry sector (Berkoz
& Turk, 2008).
Spatial Distribution of Warehouses in Istanbul
When the number of warehouses per district is examined, it can be seen that the number is
higher in suburban districts than in central districts. While the share of warehouses in
central districts is 26.7%, the share of warehouses in suburban districts is 73.3%. In
central districts, according to the number of warehouses, Eminonu has the biggest share
with 10.1%, then Sisli and Kadkoy come in with 7.7% and 6.2% shares, respectively.
In suburban districts, according to the number of warehouses, Bayrampasa has the
biggest share with 18.1%. Zeytinburnu (12.7%) and Kucukcekmece (6.2%) also have
important shares (Table 1).
The spatial distribution of warehouses has also been affected by the linear characteristics
of the city (Figure 2). Two transport axes, namely TEM (E80) and E5 (D100) highways,
have important roles in shaping the urban structure. As the city grows, E5 (D100) has
remained in the city centre over time and become an intra-city road in terms of usage.
TEM (E80) still continues to be used as a highway. The TEM E5 link roads, where
these two axes intersect, are economically developing areas and expected to develop further.
In Istanbul, warehouses were located in central districts and along the E5 (D100) ring
road until the 1990s. Together with the increase in the citys population and the expansion
of the population into the surroundings, land used for warehouse and production purposes
gained value and became more attractive for commercial and residential purposes. These
areas went through a transformation. This change over time and the construction of the
Bosphorus Bridge and the TEM (E80) ring road caused industrial areas to move farther
from the centre and outside the city (Ozus et al., 2011).
Over the years, the Ambarl port has been preferred over the Haydarpasa port, and considering locations of the Gebze Organized Industrial Zone and the Tuzla port, and since the
Ataturk Airport has now become inside the city (Figure 2), and the consideration of
directing the freight and cargo transportation to a third airport to be constructed in Istanbul,
all indicate that warehouses will move outside the city. As can be seen in Figure 3, warehouses prefer Beylikduzu, Bahcelievler, Bayrampasa, Zeytinburnu and Eminonu regions
on the European side, as they have connections to E5 as well as TEM ring roads, which
are Istanbuls two main ring roads and provide easy access to trucks and big transport
mraniye, Sargazi, Dudullu, Kartal and Tuzla regions on the
vehicles, and Samandra, U
Anatolian side, for warehousing. However, recently, as can be observed in the Figure 3,
there is a move towards the perimeter of the city. In the Gebze district attached to Izmit
near Istanbul, there is a great increase in the number of warehouses (Figure 3). The
reason for this increase is that the Gebze Organized Industrial Zone has the greatest
capacity in Turkey (Figure 2). However, it is not covered by this study since our study
only covers the province of Istanbul. Also, warehouses in the Tuzla and Ambarl
regions draw attention with their distance to ports.
The density distribution of all warehouses is provided in Figure 4. Despite the lack of
data in some districts, it is possible to understand the distribution behaviour in general

276
A. Durmus & S.S. Turk

Figure 2. Main points for warehouses in Istanbul (highways, main roads, ports, airports, CBDs, industrial zones).

Figure 3. The spatial distribution of warehouses in Istanbul.

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

277

278
A. Durmus & S.S. Turk

Figure 4. Warehouse density per district in Istanbul.

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

279

terms. In terms of warehouse density, warehouses show a highly concentrated pattern in


the traditional centre and around the traditional centre area. In particular, density values
above 150 warehouses per square kilometres in the Bayrampasa and Eminonu districts
demonstrate that there is a great warehouse stock in those districts. The reason for high
density in traditional CBDs is the accumulation of warehouses in those regions over the
years. However, as can be observed in Figure 4, there is a move towards the perimeter
of the city. In the Gebze district attached to Izmit near Istanbul, there is a great increase
in the number of warehouses.
An Analysis for Istanbul
Methodology
In the study, factors affecting the location selection of warehouse and their degree of
importance are determined in the Istanbul metropolitan area. In this study, logistic
regression is used to examine how location-specific factors influence the site preference
of warehouses. For this, with the use of GIS techniques, the development site as preferred and non-preferred groups is classified. At the metropolitan level, previous
studies used various factors in the location of warehousing, such as rents for properties,
proximity to market and production locations, highway and airport, port and rail access,
the distance to blue-collar workers and access to the central city (Sivitanidou, 1996;
Woudsma et al., 2008). In this study, factors such as transport accessibility, cluster, distance to the traditional CBD and customs, rents and market size are examined. All of
them represent the location-specific factors. The factors are defined with some explanatory
variables (Table 2).
Table 2. Explanatory variables
Variables
Accessibility
E5
TEM
Tuzlaport
Ambarlport
Zeytinburnuport
Ataturkairport
Cluster
DenIndust
DenService
Distance to city centre
Traditional CBD (Beyoglu and Eminonu)
Market size
GDP
Warehouse rents
Rents
Customs
Halkalcustoms

Definition
Distance to
Distance to
Distance to
Distance to
Distance to
Distance to

the closest E5 (D100) crossroad (km)


the closest TEM (E80) crossroad (km)
the Tuzla port (km)
the Ambarl port (km)
the Zeytinburnu port (km)
the Ataturk airport (km)

Number of industrial firms/district area (unit/km2)


Number of service sector firms/district area (unit/km2)

Distance to the old CBD (km)


GDP per district ($)
Warehouse rents monthly ($/m2)
Distance to the Halkal Customs (km)

280

A. Durmus & S.S. Turk

Hypotheses and Explanatory Variables


In the study, it is considered that the location-specific factors affect the location selection
of warehouses in the Istanbul metropolitan area. Therefore, the main hypothesis of the
study is whether the spatial distribution of warehouses in the Istanbul metropolitan area
can be explained with location-specific factors.
General hypothesis: Location-specific factors are effective for location selection of
warehouses in the Istanbul metropolitan area, and the distribution of warehouses is
not random.
Generally, it is widely accepted that transport accessibility is one of important factors
and transport plays a crucial role in the location selection of warehouses. A group of variables is used to explore the effect of accessibility to various transportation modes like
highways, ports and airports in the Istanbul metropolitan area. In Istanbul, two transportation axes, namely TEM (E80) and E5 (D100) highways, have importance for the location
selection of warehouses. As stated before, while E5 (D100) has transformed into an intracity road character over time, TEM (E80) still continues to be used as a highway.
However, the effect of two transport axes is not unknown. In the study, their effects are
tested. Also, it is considered that three ports (Tuzla Port, Ambarl Port and Zeytinburnu
Port) have importance for the location selection. Additionally, Ataturk Airport has transformed into a significant hub for national and international air traffic. Therefore, it is considered that Ataturk Airport has an effect on the location selection. The other airport of
Istanbul, Sabiha Gokcen Airport, as an independent variable has not been included in
the model since it does not create a correlation with the Tuzla Port and due to the fact
that its use for cargo purposes is very low compared with Ataturk Airport. In the study,
transport accessibility is represented by five explanatory variables. These are: E5, TEM,
Tuzlaport, Ambarlport, Zeytinburnuport and Ataturkairport (Table 2). The variables are
expected to have positive + effects on the location selection of warehouses in Istanbul.
All transportation nodes used in the study are seen in Figure 2.
Sub-hypothesis 1: Transport accessibility is effective for location selection of warehouses.
Two variables in the model are used to explore the relationship between clusters of
industrial and service firms and location selection of warehouses in the Istanbul metropolitan area. It is known that the services sector is dominant in central districts in the Istanbul
metropolitan area (Berkoz & Turk, 2008). However, in central districts, vacant land can be
hard to find and rents are very high. Therefore, these areas cannot be preferred for the
location site of warehouses. On the other hand, in suburban districts, the share of industry
firms is high. In these districts, to find vacant land can be easy and rents can be low. Therefore, these areas can be preferred for the location site of warehouses by investors in Istanbul. Considering the huge differences in the size of urban districts, clusters can be
represented by variables such as the density of industrial and services sector firms in
the urban districts. Cluster is represented by two explanatory variables in the model.
These are DenIndust and DenService (Table 2). As areas where firms in the industrial
sector are concentrated are expected to have a positive + effect for the location selection

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

281

of warehouses, areas where firms in the services sector are concentrated are expected to
have a negative 2 effect on location selection.
Sub-hypothesis 2: In regions of industrial clusters, warehouse concentration
increases, whereas in regions of clusters of the services sector, warehouse concentration decreases.
The distance to the CBD is also an important determinant as one of the location-specific
factors. In the model, the introduction of the variable is to examine both how the distance to
the CBD affects the location selection and whether there is a decentralization trend of warehouses. The Istanbul metropolitan area has a multi-central structure. Although the traditional
CBD (Beyoglu, Eminonu) has lost their importance over time, it has still the biggest share in
the number of warehouses in the central districts (Table 1). Therefore, to test the variable, the
distance to the traditional CBDs (Beyoglu, Eminonu) is used. The explanatory variable is
Traditional CBD in the model (Table 2). High rents and scarcity of vacant areas are features
of central districts. This feature is valid for the traditional CBDs. Therefore, the variable is
expected to have a negative 2 effect on the location selection.
Sub-hypothesis 3: If the distance to the traditional CBD decreases, the site preference for warehouses decreases.
The variable related to rents in the model is used to explore the relationship between
rents and location selection of warehouses in Istanbul. That is, the variable is used to
test the effect of land market on the location selection of warehouse in the Istanbul
metropolitan area. Warehouse rents are represented by Rents in the model (Table 2). In
the Istanbul metropolitan area, warehouse users are expected not to opt for areas where
rentals are high. This variable is also expected to have the negative 2 sign.
Sub-hypothesis 4: In areas where rentals are high, the site preference for warehouse
decreases.
The variable related to distance to customs in the model is used to test for the importance of customs on location selection of warehouses. In the Istanbul metropolitan area,
Halkal customs has an important function as it is very often used by logistics firms.
Customs are represented by Halkalcustoms in the model (Table 2). Therefore, the variable
is expected to have the positive + sign.
Sub-hypothesis 5: If the distance to Halkal customs decreases, the site preference
for warehouses increases.
This variable is used to assess the relationship between market size and location selection
of warehouses. As in the Istanbul metropolitan area, the districts whose GDP is higher are
preferred areas in the location selection of warehouses. Market size is represented by GDP
in the model (Table 2). This variable is expected to have the sign + in the model.
Sub-hypothesis 6: In regions whose GDP is higher, the site preference for warehouses increases.

282

A. Durmus & S.S. Turk

Data
The warehouse addresses used in the model as the dependent variable have been gathered
from surveys carried out in 2007 by the Istanbul Metropolitan Planning (IMP) operating
within the Istanbul Metropolitan Municipality (IMM). According to the 2007 IMP
records, the total number of warehouses was 7783 in Istanbul in 2007. However, this
number includes both open-air warehouses and warehouses with unclear addresses.
There are about 500 open-air warehouses registered in the IMP survey and are used for
wood, coal and scrap. This type of warehouses is not preferred for use by the real estate
sector. In the study, the warehouses that are used by the real estate sector are taken into
consideration. Therefore, open-air warehouses have not been included in the model in
order not to be misleading.
Although their numbers are known, the addresses of warehouses in the Kadkoy,
skudar, Gaziosmanpasa, Zeytinburnu, Eyup, Saryer, Sisli, Kagthane, Bagclar,
U
Pendik and Maltepe districts are not available. Despite the efforts of the researchers, the
missing data related to these districts were not acquired. In the study, the use with the
existing conditions of the official data is preferred. Therefore, those areas related to
missing data have been extracted from the model. Furthermore, there are no warehouses
located in the Sile and Adalar districts. If both open-air warehouses and warehouses with
unclear addresses are excluded, the number of warehouses decreases to 4259. This number
has been used in the model that represents 58% of the total number of warehouses (7783).
In this sense, the sample size provides opportunity to understand the distribution behaviour
of warehouses in Istanbul.
The addresses of the warehouses provided from the IMP were recorded on the GIS
program. Also, the distance to the closest E5 (D100) and TEM (E80) crossroads and
the distance to the Tuzla Port, Ambarl Port, Ataturk Airport, Zeytinburnu Port, traditional
CBD (Beyoglu, Eminonu) and Halkal Customs were calculated within the framework of
factors affecting the location sites of the warehouses. Then, with the data procured from
the IMM and the Istanbul Governorate, the concentration of firms operating in the industry
and services sectors was calculated by the district and used as data. For rents per square
metre for warehouses, the Colliers International (2010) and the Pamir and Soyuer International (2006) are used as data sources. Detailed information on data sources is given
in Table 3.

Table 3. Detailed information on data sources


Data
The addresses of warehouses
Warehouse density per quarter
The distances of warehouses to the
transportation nodes
The number of firms operating in the
industry and services sectors
Firms density in industry and services
sectors per district
Rents per square metre for warehouses

Data Sources
IMP (Istanbul Metropolitan Planning Office) 2007
Arc GIS v9.2
Arc GIS v9.2
The Istanbul Governorate (2007); IMM (Istanbul
Metropolitan Municipality) (2006)
Arc GIS v9.2
Colliers International (2010); Pamir and Soyuer
International (2006)

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

283

To measure market size, data related to GDP per district are needed. These data were
produced by TurkStat (Turkish Statistical Institute) in 1996. The data have not been
updated since this date. The using of this data can create crucial problems in the model.
Therefore, the data have not been included in the model. In 2008, Istanbul district boundaries were redrawn. However, as the data used are prior to that date, warehouse density per
district could not be constituted according to the current district settlement. All the data
have been recorded separately for each warehouse and the data set formed has been
used in the logistic regression analysis.
Model Specification
The aim of the model in the article is to break down site preference into a group of
explanatory variables and to examine how various site features contribute differently
to the preference for a site. For this, there is a need first for an indicator which is considered attractive for the region concerning logistic warehouses. Here, as a value indicating that the sites are preferred for location selection of warehouses, the warehouse
density in the quarter has been used to constitute the dependent variable. After
addresses of the warehouses have been transferred to the GIS program, warehouse
density per quarter in Istanbul has been calculated as warehouse number/km2.
Then, the average of the densities is calculated. When the average is calculated, the
extreme values are extracted in order to exactly reflect the main mass. The average
density of warehouses is found as 18 warehouses per square kilometres. Places above
this threshold value are considered in the following sections of the study as regions
where warehouses are concentrated (preferred location for warehouses), and places
below are considered to be places where there is no warehouse density (not-preferred
location for warehouses).
In the model, the transport accessibility is measured by the closest distance from the
address of each warehouse to the nodes. Some nodes are quite close each other. For
example, the distance to Sabiha Gokcen Airport and Tuzla Port and the distance to
Haydarpasa Station and Haydarpasa Port, Halkal Train Station and Halkal
Customs are nearly the same. The closeness means repetitions in the model that can
lead to a multicollinearity problem. To avoid the problem, some variables are extracted
from the model. In the model, under the transport accessibility, six dependent variables
are used. While E5 represents the distance to the closest E5 (D100) crossroad (km),
TEM represents the distance to the closest TEM (E80) crossroad (km). In the model,
Tuzlaport represents the distance to the Tuzla port (km). Ambarlport means the distance
to the Ambarl port (km). Zeytinburnuport represents the distance to the Zeytinburnu port
(km). Ataturkairport states the distance to the Ataturk airport (km). Zeytinburnuport
represents the distance to the Zeytinburnu port (km) (Table 2).
In the model, as dependent variables under cluster criteria, two dependent variables are
used. While DenIndust represents the number of industrial firms per district area (unit/
km2), DenService shows the number of service sector firms per district area. In the
model, traditional CBD (Beyoglu and Eminonu) presents the distance to the old CBD
(km). Rents demonstrate the warehouse rents monthly ($/m2) in the Istanbul metropolitan
area. Halkalcustoms represents the distance to the Halkal Customs (km) (Table 2).
Logistic regression is used to analyse relationships between a dichotomous dependent
variable and metric or dichotomous independent variables. The logistic model is widely

284

A. Durmus & S.S. Turk

used in economics and market research (McFadden, 1978; Shukla & Waddell, 1991; Wu,
2000). The distribution of warehouses demonstrates an uneven pattern in the Istanbul
metropolitan area. To explain whether a site is a preferred location for a logistic warehouse
in the Istanbul metropolitan area, the preferred sites or not-preferred sites as a
dichotomous dependent variable are used. A dichotomous dependent variable can easily
be explained with the logistic regression.
Here, preference of a site will depend on a set of variables that includes location-specific
factors in the location selection of warehouses. If, for example, a warehouse investment
in the preferred sites is in Istanbul, the dependent variable z takes the value of 1 for
preferred sites, and the value of 0 for being not-preferred sites in the Istanbul metropolitan area. The logistic model performs a maximum likelihood estimation of models with
dependent variables coded as 0/1. Warehouses with density above 18, within the preferred
region, are coded 1, those under 18, which are not within the preferred region, are
coded 0

z=


1 areas where concentration is observed(preferred location for warehouse)
.
0 areas where there is no concentration is observed(non preferred location for warehouse)

Vector z represents the composite preference of a site. The purpose is now to break
down z into linear combination variables x1, x2, x3, . . ., xp which are site attributes:
z = b0 + b1 x1 + b2 x2 + b3 x3 + + bp xp .

(1)

Here, the value z indicates at the same time the ratio of the probability of the region to be
preferred to the probability of it being not preferred. In case this non-linear relationship is
established, the logistic transformation is adopted. Thus, z is defined as a logit, which is a
log ratio of the probability of the region to be frequently preferred to the probability of it
being not preferred
z = logit = ln

 a
p
pb

(2)

where pa and pb are, respectively, the probability of the site to be frequently preferred
and the probability of it being not preferred.
Since
pa + pb = 1
and hence:

z = ln

pa
1 pa


(3)

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

285

If pa denotes the probability of being a preferred site:


pa =

exp(z)
1 + exp(z)

(4)

The logistic estimate provides information about which of the characteristics included
in vector z plays an important role in the location selection of warehouses.

Empirical Findings
In a logistic regression model, there are two indexes to describe the global fit of the model
that represent the proportion of uncertainty of the data explained by the adjusted model.
Through analogy with the determination coefficient in the linear regression, they are
represented by R2 corresponding to the R2 of Cox and Snell and the R2 of Nagelkerke.
The value of the R2 of Cox and Snell has the inconvenience of not reaching the value 1
(100%) when the model reproduces the data exactly (Domnguez-Almendros et al.,
2011, p. 5).
Another measure that describes the global fit of the model is the chi-squared goodnessof-fit test. This is a chi-squared goodness-of-fit statistic that compares the observed values
with the values predicted by the model. Since the model has been estimated with the
maximum likelihood method, its global significance, i.e. the significance of the set of
included predictor variables, is assessed with the so-called likelihood ratio test. The
likelihood ratio test for studying the significance of the model involves comparing the
goodness of fit of the saturated model (including all the variables) with the null model
(adjusted by a constant) through the deviation ratio (deviance 22 log (likelihood)).
A statistic is constructed that follows a chi-squared distribution with (number of variables
of the saturated model21) degrees of freedom (df) (Domnguez-Almendros et al., 2011,
p. 5).
Table 4 shows the results of the logistic regression in summary. In the model, 2
2 log(likelihood) is measured as 1871.672. Cox and Snells R2 is 0.509. Nagelkerkes
R2 is normally higher than the Cox and Snell measure. In the model, it is 0.745, indicating
a moderately strong relationship of 74.5% between the predictors and the prediction
(Table 4) Also, the chi-squared goodness-of-fit test is found statistically significant
(x2 3031.179, Sig. 0.000) (Table 5).
When the classification table is examined, 3094 of 3141 regions preferred for warehouses (98.5% of those) and 853 of 1118 regions not preferred (76.3% of those) have
been estimated correctly with the model. In conclusion, 92.7% of 4259 preferences
were estimated correctly (Table 6).
According to the results of the model (Table 7), it is found that transport accessibility
(E5, TEM, Tuzlaport, Ambarlport, Zeytinburnuport and Ataturkairport), cluster
Table 4. Model summary
Step
1

22 log likelihood

Cox and Snells R2

Nagelkerke R2

1871.672

0.509

0.745

286

A. Durmus & S.S. Turk


Table 5. Omnibus tests of model coefficients

Step
Block
Model

x2

df

Sig.

3031.179
3031.179
3031.179

11
11
11

0
0
0

(DenIndust, DenService), distance to the city centre (traditional CBD), warehouse rents
(rents) and customs (Halkalcustoms) are statistically significant in the location selection
of warehouses in the Istanbul metropolitan area. This means that location-specific factors
are effective for location selection. The expectations related to sub-hypotheses and their
results are summarized in Table 8.
According to sub-hypothesis 1, the transport accessibility is expected to have a favourable impact on location selection of warehouses. When the results are examined, it appears
that this hypothesis is proved. Warehouses are sensitive to transport accessibility.
According to the model, Ataturk Airport and Zeytinburnu Port are in a leading position
in variables related to transportation accessibility. The fact that Ataturk Airport has a
47% cargo share naturally increases its importance for the location site of warehouses.
However, in the model, the negative sign indicates that along with a decrease in distance
to the Ataturk Airport, the site preference for warehouses decreases. This can be explained
that land prices in the close surroundings of the airport are very high and the finding of
vacant land is difficult. Also, Zeytinburnu Port exerts a consistently strong influence on
the location site of warehouses. The positive sign of Zeytinburnu Port demonstrates that
along with a decrease in distance to Zeytinburnu Port, the site preference for warehouses
increases. This can be explained with being in an easy accessible location out of Haydarpasa Port in central districts and an old industrial zone. The coefficient for Zeytinburnu
Port is +4.376. In this case, exp (+4.376) 1.476. This means that with every decrease
of 1 km distance to Zeytinburnu Port, the odds of preferred regions to not-preferred regions
increase by about 147.6%.
The distance to Ambarl Port is one of the indispensable criteria for warehouses when
compared with other ports of the city. However, the only disadvantage of the region is that
warehouse zoned lands in its surroundings are exhausted or have higher prices. Otherwise,
it is definitely a centre of attraction. In the model, the negative sign indicates that along
with an increase in distance to Ambarl Port, the site preference for warehouses increases.
Also, the distance to Tuzla Port is another important accessibility variable. Compared with
Ambarl Port, it is more advantageous in terms of vacant land. However, in the model, the
Table 6. Classification table
Predicted
Observed
0
1
Overall percentage
The cut value is 500.

0
853
47

1
265
3094

Percentage correct
76.30
98.50
92.70

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

287

Table 7. Results of logistic regression


Dependent variable: site preference
Preferred location for warehouse: 1
Non-preferred location for warehouse: 0
2 log likelihood 1871.672
Variables
Tuzlaport
Zeytinburnuport
Ambarlport
Ataturkairport
Halkalcustoms
E5
TEM
Traditional CBD
DenIndust
DenService
Rents
Constant

Coefficients (B)
20.318
1.476
20.265
21.475
0.477
0.380
20.133
21.022
0.014
20.023
22.118
35.534

Wald
24.540
163.750
4.687
148.699
19.819
29.540
3.204
161.464
63.376
187.826
71.037
50.653

Exp (B)
0.727
4.376
0.767
0.229
1.611
1.462
0.876
0.360
1.014
0.977
0.120
0.000027

Sig.

Significant at the 0.01 level.


Significant at the 0.05 level.

Significant at the 0.10 level.

variable has a negative sign. The negative sign can be explained by the fact that Tuzla Port
is located near the Gebze Organized Industrial Zone. Therefore, this port is used much
more by the near surroundings and outside of the Istanbul metropolitan area borders.
Table 8. The expected effects of variables in the model and findings
Variables

The expected effects

The findings

Accessibility
E5
TEM
Tuzlaport
Ambarlport
Zeytinburnuport

+
+
+
+
+

+
2
2
2
+

Cluster
DenIndust
DenService

+
2

+
2

Distance to city centre


Traditional CBD (Beyoglu and Eminonu)

Warehouse rents
Rents

Customs
Halkalcustoms

288

A. Durmus & S.S. Turk

According to the results of the model, the effects of TEM (E80) and E5 (D100) highways on the site location of warehouses are different. While the distance to E5 (D100)
increases the site preference, the distance to TEM (E80) does not. This difference can
be explained by their functions in the Istanbul metropolitan area. E5 (D100) has been
transformed into an intra-city road character over time, while TEM (E80) still continues
to be used as a highway. That is, according to the results of the model, warehouses are
much more sensitive to an intra-city road character than to highways in Istanbul.
According to sub-hypothesis 2, as areas where firms in the industrial sector are concentrated are expected to have a positive + effect for the location selection of warehouses,
areas where firms in the services sector are concentrated are expected to have a negative
2 effect on location selection. According to the results of the model, it appears that the
relationship between industry and services clusters and the choice of location is statistically significant. That is, sub-hypothesis 2 is proved. It is observed that industrial clusters
have a favourable impact on location selection, whereas the services sector has an
unfavourable impact. Considering in particular the Ikitelli, Hadmkoy,Tuzla, Dudullu
and Beylikduzu industrial zones (Figure 2), it is quite normal that warehouses are
concentrated near these places. Similarly, the situation is the opposite in the CBDs
where the services sector is predominant.
In sub-hypothesis 3, the distance to the traditional CBD has an expected negative 2
effect on warehouse location selection. Sub-hypothesis 3 has been proved by the model.
The variable is found to have a negative 2 effect on location selection. In the model,
the negative sign indicates that along with the decrease in distance to the traditional
CBD (Beyoglu, Eminonu), the site preference for warehouses decreases. This is not
surprise result. Within the traditional CBD, although there is better access to services,
vacant land can be hard to find. Due to disadvantages such as traffic density, high rents
and being far from industrial clusters, the investors do not prefer to be located in the
central business areas.
In sub-hypothesis 4, warehouse users are expected not to opt for areas where rentals are
high. This variable is also expected to have the negative 2 sign. According to the results
of the model, the relationship between warehouse rents and location selection has been
found statistically significant. That is, there is an effect of land markets on location
selection. In the model, the negative sign indicates that along with increase in rents, the
site preference for warehousing decreases. In this case, exp (20.120) 22.118. This
means that for every increase of US$1 per square metres for warehouse rents monthly,
the odds of preferred districts to the not-preferred districts decrease by about 211.8%.
According to sub-hypothesis 5, the distance to Halkal Customs is expected to have a
positive + sign. This variable is found statistically significant. That is, sub-hypothesis
5 has been proved by the model. The positive sign of Halkal Customs shows that along
with a decrease in the distance to Halkal Customs, the site preference for warehousing
increases. In this case, exp (+1.611) 0.477. This means that for every increase of
1 km distance to Halkal Customs, the odds of preferred districts to the not-preferred
districts increase by about 47.7%.
General Evaluation and Conclusion
Warehousing establishments are concentrated in metropolitan areas. However, how logistic warehouses are distributed or what their logic in location is at the intra-urban level is

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

289

not known exactly. In the literature, the empirical studies on location selection of warehouses are very limited at the intra-urban level. In Turkey and specifically in the Istanbul
metropolitan area, the need for warehouses has considerably increased. Warehouses have
started to be used by the real estate sector. In this sense, location selection of warehouses
can affect the structure and form of the city.
In this article, factors affecting the location selection of warehouses and their degree of
importance are determined by using a logistic regression model focusing on the Istanbul
metropolitan area. As a dependent variable, the distinction of regions preferred and nonpreferred in location selection is used. In the model, within the framework of factors
affecting location selection transport accessibility (E5, TEM, Tuzlaport, Ambarlport,
Zeytinburnuport and Ataturkairport), cluster (DenIndust and DenService), warehouse
rents (rents), the distance to the city centre (traditional CBD) and customs (Halkalcustoms) are used as independent variables. It was found that location-specific factors are
effective for the location selection of warehouses in the Istanbul metropolitan area.
That is, the distribution of warehouses is not random, and warehouse location follows a
certain economic rationality at the intra-urban level.
According to the results of Sivitanidous (1996) study for Los Angeles, higher rents for
properties are sensitive to highway and airport access, but not port or rail access. Additionally, in the study of Woudsma et al. (2008), access to highways and the airport led to more
logistics development. Different from the findings of these studies, in the Istanbul metropolitan area, warehouses are sensitive to all transport nodes. However, Ataturk Airport and
Zeytinburnu Port are in a leading position related to transportation accessibility.
However, warehouse investors cannot be located near Ataturk Airport because the land
prices in the close surroundings of the airport are high. Also, the other ports like Tuzla
Port and Ambarl Port have different effects on the site preference for warehouses. The
reasons for this are that Ambarl Port has a disadvantage of the region that warehouse
zoned lands in its surroundings are exhausted or have higher prices. Therefore, the investors
cannot prefer this area for location selection of warehouse. Although Tuzla Port has an
advantage in terms of vacant land, compared with Ambarl Port, this port is much more
used by the close surroundings outside of Istanbul metropolitan area borders. The other
interesting result in the model is that warehouses are much more sensitive to an intra-city
road character than to highways in the Istanbul metropolitan area.
According to the model, industrial clusters made a favourable impact on the location
selection of warehouses; however, services sector clusters made an unfavourable
impact. In the Istanbul metropolitan area where the services sector is concentrated, it is
nearly impossible to find warehouses at affordable prices. Furthermore, in these regions,
because the finding of vacant land is hard, a warehouse project proposal cannot be feasible.
In the model of Woudsma et al. (2008), access to the central city was found irrelevant.
According to the model, warehouse investors prefer to be located in the periphery of the
city instead of in the traditional CBD. That is, it can be said that there is a decentralization
tendency from the traditional CBD to the periphery in the Istanbul metropolitan area. The
result is supported by warehouse rents. According to the results of the model, the land
market has an impact on the location selection of warehouses. The other important
result is that Halkal Customs is a crucial point for the distribution of warehouses. In
the Istanbul metropolitan area, in the future, after completion of the Gumuskaya Port
and the Marmaray Project could lead to important changes in the logistics sector.
Therefore, these factors affecting location selection can be changed.

290

A. Durmus & S.S. Turk

In the future studies, data on warehouses in the C


orlu and Gebze regions can be used
to assess the impact on the selection of warehouse location in neighbouring provinces
and/or other transport means such as airports. These regions can be added to the model
as criteria for location selection.

References
Alberto, P. (2000) The logistics of industrial location decisions: An application of the analytical hierarchy process
methodology, International Journal of Logistics: Research and Application, 3(3), pp. 273289.
Babacan, M. (2005) Lojistik sektorunun ulkemizdeki gelisimi ve rekabet vizyonu [The development and competition vision of logistic sector in Turkey], Ege Akademik Baks, 5(1), pp. 916, (in Turkish).
Berkoz, L. & Turk, S. S. (2008) Determination of location-specific factors at the intra-metropolitan level: Istanbul
case, Tijdschrift voor Economische en Sociale Geografie, 99(1), pp. 94114.
Bodenam, J. (1998) The suburbanization of the institutional investment advisory industry: Metropolitan
Philadelphia, 1983 1993, Professional Geographer, 50(1), pp. 112126.
Bowen, J. T. (2008) Moving places: The geography of warehousing in the US, Journal of Transport Geography,
16(6), pp. 379387.
Cidell, J. (2010) Concentration and decentralization: The new geography of freight distribution in US metropolitan areas, Journal of Transport Geography, 18(3), pp. 363371.
Coffey, W. J., Drolet, R. & Pole`se, M. (1996) The intra-metropolitan location of high order services: Patterns,
factors and mobility in Montreal, Papers in Regional Science, 75(3), pp. 293323.
Colliers International (2010) Turkiye Real Estate Review (Istanbul: Colliers International).
Demirel, T., Demirel, N. C. & Kahraman, C. (2010) Multi-criteria warehouse location selection using Choquet
integral, Expert Systems with Applications, 37(5), pp. 39433952.
Dogan, M. (2006) Isletme Ekonomisi ve Yonetimi [Management economy and administration], No. 57 (Izmir:
Birlesik Matbaaclk) (in Turkish).
Domnguez-Almendros, S., Benitez-Parejo, N. & Gonzalez-Ramirez, A. R. (2011) Logistic regression models,
Allergologia et Immunopathologia, 39(5), pp. 295305.
Drezner, Z., Scott, C. & Song, J. S. (2003) The central location problem revisited, IMA Journal of Management
Mathematics, 14(4), pp. 321336.
Durmus, A. (2010) Lojistikte depo yer secimine etki eden faktorlerin modellenmesi [The modelling of factors
affect the location choice of warehouse: Istanbul case]. Unpublished MSc thesis, Istanbul Technical University, Graduate School of Science Engineering and Technology, Istanbul.
Eaton, L. (1982) Warehouses and warehouse districts in mid-American cities, Urban History Review, 10(3),
pp. 1726.
Eilon, S. (1982) Multi criteria warehouse location, International Journal of Physical Distribution and Logistic
Management, 12(1), pp. 4245.
Ersoy, M. S. (2006) Lojistik ve Turkiyenin konumu [Logistic and the location of Turkey], Galatasaray
niversitesi Iktisadi ve Idari Bilimler Fakultesi, Durum Dergisi, 4(1), pp. 2429.
U
Fulton, M. (1971) New factors in plant location, Harvard Business Review, 6(5), pp. 411.
Gezici, F., Giritlioglu, C., Ocakci, M., Yuzer, M. A., Ayatac, H. & Oruc, G. D. (2008) Evaluation of cluster
behavior of the manufacturing sector in the Istanbul Metropolitan Area. Regional Science Association
International (RSAI) World Congress, Sao Paulo, Brazil.
Glasmeier, A. & Kibler, J. (1996) Power shift: The rising control of distributors and retailers in the supply chain
for manufactured goods, Urban Geography, 17(8), pp. 740757.
Gurdal, S. (2006) Turkiye Lojistik Sektoru Altyap Analizi [Infrastructure Analysis of Logistic Sector in Turkey].
Istanbul Ticaret Odas Yayn, No. 200614, Istanbul (in Turkish).
Heikkila, E., Gordon, P. & Kim, J. I. (1989) What happened to the CBD distance gradient? Land values in a
polycentric city, Environment and Planning A, 21(2), pp. 221232.
Hong, J. (2007) Firm-specific effects on location decisions of foreign direct investment in Chinas logistic
industry, Regional Studies, 41(5), pp. 673683.
IMM (2006) Sanayi Sektoru Analizleri Raporu [Analysis Report for Industry Sector]. Istanbul: Istanbul Buyuksehir Belediyesi Sehir Planlama Mudurlugu (in Turkish).

Factors Influencing Location Selection of Warehouses at the Intra-Urban Level

291

IMM (2008) Istanbul Hizmet Sektoru Analiz Calsmalar Sonuc Raporu [Final Analysis Report Related to Service
Sector in Istanbul]. Istanbul: Istanbul Metropolitan Municipality, Urban Planning Department, January 2008
(in Turkish).
IMP (2006a) Istanbulun Kentsel Lojistik Planlamas [Urban Logistic Planning of Istanbul]. Istanbul: Istanbul
Metropolitan Municipality, November 2006 (in Turkish).
IMP (2006b) C
evre Duzeni Plan Raporu [Report of Superior Land Use Plan], pp. 427 471. Istanbul: Lojistik
Hizmetler Yaps (in Turkish).
IMP (2006c) C
evre Duzeni Plan Sentez Raporu [Syntheses Report of Superior Land Use Plan], pp. 171 199.
Istanbul: Lojistik Hizmetler Sentezi (in Turkish).
IMP (2006d) Istanbul Metropoliten Alannda Sanayi Sektoru [Industry Sector in Istanbul Metropolitan Area].,
November 2006 (Istanbul: Istanbul Metropolitan Municipality) (in Turkish).
Istanbul Governorate (2007) Istanbul istihdam raporu [Istanbul Employment Report], Istanbul: Istanbul
Governorate (in Turkish).
Kahraman, C., Cebeci, U. & Ulukan, Z. (2003) Multi-criteria supplier selection using fuzzy AHP, Logistics
Information Management, 16(6), pp. 382394.
Korpela, J. & Touominen, M. (1996) A decision aid in warehouse site selection, International Journal of
Production Economics, 45(13), pp. 169180.
Korpela, J., Lehmusvaara, A. & Nisonen, J. (2007) Warehouse operator selection by combining AHP and DEA
methodologies, International Journal of Production Economics, 108(12), pp. 135142.
Maccarty, B. L. & Atthrirawong, W. (2003) Factors effecting location decisions in international operations A
Delphi study, International Journal of Operations and Production Management, 23(7), pp. 794818.
Mcfadden, D. (1978) Modelling the choice of residential location, in: A. Karlquist, L. Lundquist, F. Snickars
& J. L. Weibull (Eds) Spatial Interaction Theory and Planning Models, pp. 7596 (Amsterdam: NorthHolland).
Mckinnon, A. (2009) The present and future land requirements of logistical activities, Land Use Policy, 26(Supp.
1), pp. 293301.
Melo, M. T., Nickel, S. & Saldanhada-Gama, F. (2009) Facility location and supply chain management: A review,
European Journal of Operational Research, 196(2), pp. 401412.
Ministry of Information, Industry and Technology (2009) Istanbulun sanayi profili [Profile of Industry in
Istanbul], Report, Ankara: Ministry of Information, Industry and Technology.
Ozcan, T., Celebi, N. & Esnaf, S. (2011) Comparative analysis of multi-criteria decision making methodologies
and implementation of a warehouse location selection problem, Expert Systems wth Applications, 38(8),
pp. 97739779.
Ozdemir Darby, D. (2008), Lojistik Merkez Olarak Istanbulun Degerlendirilmesi [Evaluation of Istanbul as a
niversitesi Muhendislik Mimarlk Fakultesi) (in Turkish).
Logistic Center] (Istanbul: Yeditepe U
Ozdemir Darby, D. (2010) Strategic choice for Istanbul: A domestic or international orientation for logistics?
Cities, 27(3), pp. 154163.
Ozus, E., Turk, S. S. & Dokmeci, V. (2011) Urban restructuring of Istanbul, European Planning Studies, 19(2),
pp. 331356.
Pamir and Soyuer International. (2006) Property Advisors Report, Istanbul: Pamir and Soyuer International Co.
Pole`se, M. & Shearmur, R. (2004) Is distance really dead? Comparing industiral location patterns over time in
Canada, International Regional Science Review, 24(4), pp. 431 457.
Rodrigues, A., Bowersox, D. J. & Calantone, R. J. (2005) Estimation of global and national logistics expenditure:
2202 data update, Journal of Business Logistics, 26(2), pp. 115.
Shearmur, R. & Alvergne, C. (2002) Intra-metropolitan patterns of high-order business service location, a comparative study of seventeen sectors in Ile-de-France, Urban Studies, 39(7), pp. 11431163.
Shearmur, R. & Coffey, W. J. (2002) A tale of four cities: Intra-metropolitan employment distribution in Toronto,
Montreal, Vancouver and Ottawa-Hull, 1981 1996, Environment and Planning A, 34(4), pp. 575598.
Shukla, V. & Waddell, P. (1991) Firm location and land use in discrete urban space: A study of the spatial structure of Dallas-Fort worth, Regional Science and Urban Economics, 21(2), pp. 225253.
Sivitanidou, R. (1996) Warehouse and distribution facilities and community attributes: An empirical study,
Environment and Planning A, 28(4), pp. 12611278.
Tabari, M., Kaboli, A., Aryanezhad, M. B., Shahanaghi, K. & Siadat, A. (2008) A new method for location
selection: A hybrid analysis, Applied Mathematics and Computation, 206(2), pp. 598606.
BITAK (2006), Lojistik Arastrma Raporu [Logistic Research Report]. Project no: 35, Tubitak & Ulastrma
TU
Bakanlg [Ministry of Transport], Ankara (in Turkish).

292

A. Durmus & S.S. Turk

TURKSTAT (2010) Turkeys Statistical Year Book (Ankara: Turkish Statistical Institute).
Vlachopoulou, M., Silleos, G. & Manthou, V. (2001) Geographical information systems in warehouse site
selection decisions, International Journal of Production Economics, 71(1), pp. 205212.
Woudsma, C., Jensen, J., Kanaroglou, P. & Maoh, H. (2008) Logistics land use and the city: A spatialtemporal
modeling approach, Transportation Research Part E, 44(2), pp. 277297.
Wu, F. (2000) Modelling intra-metropolitan location of foreign investment firms in a Chinese city, Urban Studies,
37(13), pp. 24412464.
Wu, J. & Radbone, I. (2005) Global integration and the intra-urban determinants of foreign direct investment in
Shanghai, Cities, 22(4), pp. 275286.
Standard & Poors (2011) RatingsDirect on the Global Credit Portal, December 5, 2011.
Datamonitor (2009) Global top 10 Logistics companies Industry, Financial and SWOT Analysis, August 2009,
p.119. Available at http://www.researchandmarkets.com/reports/1088195/global_top_10_logistics_companies_industry (accessed 6 August 2012).

Copyright of European Planning Studies is the property of Routledge and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.

You might also like