Transport Policy 44 (2015) 3747

Contents lists available at ScienceDirect

Transport Policy
journal homepage: www.elsevier.com/locate/tranpol

How logistics performance of freight operators is affected by urban

freight distribution issues
Jos Geraldo Vidal Vieira a,n, Jan C. Fransoo b
a
b

Department of Production EngineeringDEPS, Rodovia Joo Leme dos Santos (SP-264), Km 110, Bairro Itinga, CEP 18052-780 Sorocaba, SP, Brazil
School of Industrial Engineering, P.O. Box 513, Pav F4, NL-5600 MB Eindhoven, Netherlands

art ic l e i nf o
Article history:
Received 9 October 2014
Received in revised form
3 June 2015
Accepted 22 June 2015
Keywords:
Regulations
Collaboration
Logistical performance
Detour

a b s t r a c t
This article is to shed light on the interactions among the various freight distribution constructs such as
regulations, collaboration, detour, load/unload interfaces and logistical performance. The proposed model is
empirically tested using Partial Least Squares with 119 freight operators. The ndings reveal the moderating
effect of regulations (negative effect) on the positive relationship between collaboration and load/unload
interfaces regarding receivers and freight operators. According to the effects shown by our model, regulation,
along with lack of collaboration, appear to be the Achilles' heel of freight distributors, in that both factors
contribute (directly and indirectly) to detour, which results in less efcient logistical performance.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction
Economic growth in metropolitan urban areas is resulting in increasing demands from private and public companies for goods and
services (Cherrett et al., 2012). As a result, demand has risen for freight
transportation services and land use in dense urban areas. Moreover,
freight activities to and from warehousing and distribution facilities in
these areas have become increasingly decentralised (Cidell, 2010),
heightening concerns about the reliability and maintenance of urban
infrastructure. Urban freight management is often hampered by a high
number of stakeholders (Stathopoulos et al., 2012), each focusing on
their own activities to the detriment of systematic comprehension of
interactions among several constructs (e.g., land use, type of goods
moved, nature of transfer operations, relationship between goods
type, truck type and operating partners, and effect of regulations)
(Woudsma, 2001). Most delivery occurs without a full understanding
of the freight distribution system (Nuzzolo and Comi, 2014) with regard to suitable vehicle parking, loading, and unloading. From a
strategic point of view, there is no joint planning among the distribution-chain stakeholders. Issues such as trafc congestion, thefts,
and truck restriction have increased over the past two decades and
challenged truck operators in the freight distribution system. For example, the reduction of stockholding space in urban shops, ofces, and
factories has resulted in increased demand for reliable, regular, exible
n

Corresponding author.
E-mail addresses: jose-vidal@ufscar.br (J.G. Vidal Vieira),
J.C.Fransoo@tue.nl (J.C. Fransoo).
http://dx.doi.org/10.1016/j.tranpol.2015.06.007
0967-070X/& 2015 Elsevier Ltd. All rights reserved.

delivery of relatively small quantities of products (De Vasconcellos,

2005). However, retailers may not consider that suitable parking and
unloading areas, and personnel to unload or examine the vehicles/
goods at the receiving establishment, are required.
Two factors may result in inefcient freight distribution systems in dense urban areas. First, the lack of collaboration between
logistics service operators and receivers has indirectly contributed
to such problems as trafc congestion and use of unsuitable unloading and parking areas. For example, without information or
exibility to deal with contingencies, operators may cause trafc
congestion when delivering goods by using unsuitable parking
areas to unload. Second, governmental regulations aimed at decreasing trafc congestion, harmonising land use, reducing environmental burdens, and improving logistical ow can directly
and negatively affect the freight distribution system by exacerbating imbalance between use of light goods vehicles and land
availability for parking, loading/unloading areas, and trafc safety.
Freight distribution issues are sometimes affected more negatively
by regulations than by lack of collaboration between the companies. Moreover, regulations have an indirect negative impact on
rms' logistical performances: by increasing the number of vehicles circulating in central areas and restricting truck circulation
(Thompson and Taniguchi, 2001), regulations can create more
trafc congestion (De Vasconcellos, 2005) and necessitate the use
of narrow streets and unsafe areas to deliver goods. In turn, these
factors increase delivery time and the range of vehicles needed,
and make it more difcult to schedule alternate routes.
The objective of our article is to shed light on the interactions

38

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

among the various freight distribution constructs related to trafc

congestion, urban infrastructure and its regulations, collaboration
between partners, logistical delivery aspects of goods, and rms logistical performance constructs. Specically, our research highlights
the direct, indirect, and moderating effects of governmental regulations, as well as the lack of collaboration among rms. To our
knowledge, there is no study of the interaction among the freight
distribution constructs, and literature that examines the impact of
the regulation, lack of collaboration, detour and load/unload interfaces (e.g., lack of parking, loading/unloading areas) on distribution
performance of the companies (Anderson et al., 2005; Hesse and
Rodrigue, 2004) is limited. Therefore, a contribution of our research
is to provide a model that measures these impacts according to rm
size. We also investigate the moderating effects of regulations on the
relationship between collaboration and load/unload interfaces
among the companies, and add to the negligible amount of literature
on the importance of in-depth regulations analysis to the improvement of company logistical performance. We use the So Paulo
Metropolitan Region (SPMR), the sixth largest metropolitan region in
the world, as an example of a megacity that presents many urban
freight distribution issues. We believe the lessons of the research
results extend to other megacities and large cities.
In this article, based on theoretical background, we develop separate hypotheses for specic relationships between constructs, with
regard to the main constructs. Next, we outline our methodology,
including the framework of constructs and respective indicators. We
follow with empirical data analysis and results, including a measurement model and a structural model, and present our discussion and
conclusions, including theoretical, managerial and governmental implications, study limitations and opportunities for further research.

2. Theoretical background and hypotheses

2.1. Effects of regulations
Municipalities have implemented trafc regulations in transportation systems to reduce vehicle congestion and improve mobility
and trafc organisation in large centres (Dablanc, 2007). However,
they typically address such easy-to-implement factors as vehicle
weights and time zones (Behrends et al., 2008). Although several
cities have successfully mitigated trafc congestion by implementing
truck restrictions, negative externalities have emerged from this
controversial policy. Woxenius (2012) argues regulations add restriction zones that can cause detour, the term we use to refer to a
result of trafc congestion, narrow streets and unsafe delivery areas.
For example, trafc congestion may be increased when several
competitors serve less-than-truckload markets where there are truck
restriction zones. Castro and Kuse (2003) report regulations such as
truck restriction zones and restricted timetable circulation signicantly increase total vehicle-kilometres, vehicle-hours, and use of
many smaller vehicles to complete the deliveries, in addition to increasing trafc congestion (Lindholm and Behrends, 2012). These
restrictions also contribute to an increase in use of unsafe areas and
alternate routes that have physical constraints such as narrow road
widths and low vertical clearances (Castro and Kuse, 2005). According to Holgun-Veras (2010), restrictions are frequently placed on
the use of large trucks without considering that trucks are more efcient than smaller vehicles. Therefore, large truck restrictions can
ultimately increase smaller truck trafc and result in more externalities than those produced by large trucks. The author illustrates
this concept with two important externalities: pavement deterioration and road space consumption. Thus, we formulate the following
hypothesis:

H1. Regulation is positively related to detour.

In addition, strict timetables and zone circulation for cargo vehicles
restrict availability of parking and unloading areas and contribute to
an increase in long queue, which can in turn cause detour. Munuzuri
et al. (2005) argue that due to limited time windows and lack of
suitable parking and unloading areas (Cherrett et al., 2012), freight
vehicles are often forced to enter congested areas during peak hours,
thus worsening trafc detour and resulting in more delivery delays,
increased theft, and increased use of many small delivery vehicles.
Therefore, we derive the following two hypotheses:
H1a. Load/unload interfaces mediate the positive effect of regulation on detour.
H1b. Detour mediates the positive effect of regulation on logistical
performance.
2.2. Effects of collaboration on relationship between logistical service
operators and receivers
A high level of coordination is required in an integrated freight
distribution system (Hesse and Rodrigue, 2004). Many companies may
be connected to one another to deliver goods to end users; they must
also address trafc issues, signicant demand increases, high-frequency deliveries in lighter vehicles, and deliveries in risky urban
areas (Crainic et al., 2004). Our research focuses on collaboration aspects that involve sharing logistical information, exibility, and commitment between freight operators and retailers during goods delivery. Sharing logistical information is important because not all companies use information technology. Increased information sharing in
logistical operations may solve logistical contingency issues among
the partners (e.g., availability of loading and unloading areas, parking
areas). Flexibility is also important: for example, rms could arrange to
deliver goods at off times or provide dedicated logistical operation
time to delivery for key accounts. Furthermore, relationships based on
commitment between partners (Morgan and Hunt, 1994) may lead to
service level improvements and reduce order cycle times and inventory levels (Moberg et al., 2002).
Characteristics of the relationships between operators and receivers have strong impacts on the delivery decision in terms of
aggregate times and costs; operators are more likely to aggregate
costs alone if there is relatively less time available to satisfy the
delivery (Puckett and Hensher, 2009). In this case, the stricter time
window exerts pressure on the retailer to be exible on when
goods are received. However, retailers incur greater costs from offpeak delivery than operators (Holguin-veras et al., 2005). As a
result, shippers and operators can signicantly inuence delivery
times, while receivers have little input into when the vehicles arrive (Cherrett et al., 2012). Truck operators focus on fast delivery of
goods and demand better accessibility (mainly off peak hours),
trafc information, and control of loading/unloading bays (Stathopoulos et al., 2012); they generally prefer routes with many
stops over a single delivery (Figliozzi, 2010). In contrast, receivers
prefer to receive goods during the day (Domnguez, et al., 2012).
Therefore, a retailer's lack of temporal exibility and reluctance to
receive goods in peak delivery due to daily commercial activities,
combined with the shipper's failure to divulge delivery time, may
result in operators parking and unloading vehicles in unsuitable
places. Therefore, we derive the following hypothesis:
H2. The lack of collaboration is positively related to load/unload
interfaces.
Detour is also indirectly inuenced by a lack of collaboration
between partners. Because partners have different interests in
freight distribution systems, related to time and local delivery issues, they are not willing to collaborate with one another, assume

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

their own costs, and handle problems independently. Their choice

of the shortest path or most convenient location to park and unload the goods may increase total detour in a city. Stathopoulos
et al. (2012) show restricted time windows and a lack of parking
and loading/unloading bays are the most severe problems for
traders and freight operators. Real-time trafc information is also
important for them. Thus, we argue the following:
H2a. The load/unload interfaces mediate the positive effect of lack
of collaboration on detour.
Literature indicates there is little collaboration between companies
in freight distribution (Lindholm and Behrends, 2012), resulting in
unsuitable use of land for parking and unloading and lower logistic
performance (Castro and Kuse, 2005) through delivery delays, thefts
due to parking in unsuitable areas, and use of unsafe areas to unload.
Therefore, we derive the following hypothesis:
H2b. Load/unload interfaces mediate the negative effect of lack of
collaboration on logistical performance.

39

loading, and unloading areas are required, resulting in problems of

trafc congestion and use of narrow streets and unsafe areas. In dense
urban areas, lack of space and increased deliveries result in shippers
using unsuitable spaces for delivery vehicle parking (Munuzuri et al.,
2005). According to Stathopoulos et al. (2012), operators have suggested that use of reserved lanes for goods distribution would be a
useful policy for handling restrictive delivery time windows, and decreasing detour.
Long queues and unsuitable loading/unloading and parking
areas cause delays, which disrupt route planning and result in
more time spent on delivery (Munuzuri et al., 2005). Due to a lack
of suitable receiver places to unload some vehicle sizes, the
availability of vehicles is reduced. Cargo theft may also increase
due to the distance from the goods vehicle to the premises being
served; the company is forced to park in another location, and the
driver has to close and lock the vehicle or pay a local assistant to
examine the vehicles and goods.
Based on these factors, we hypothesize the following:
H4. Load/unload interfaces are positively related to detour.

2.3. The moderating effect of regulations

A moderator is a construct that affects the strength of a causal
relationship between two other constructs (Lowry and Gaskin, 2014).
In our research, we nd regulations inversely strengthen the relationship between collaboration and load/unload interfaces: stricter
regulations lessen the load/unload interfaces presented by lack of
collaboration. In other words, the lack of suitable parking and loading/
unloading areas may be less affected by the lack of information
sharing, commitment, and exibility between the actors when zone
restrictions and restricted truck timetables are in effect. According to
Castro and Kuse (2005), truck restriction schedules are unfavourable
to some companies because they result in fragmented working hours
for their drivers and other personnel, thus hampering the relationship
between partners. Furthermore, traditional infrastructure and regulatory policy cannot satisfy current and future demand levels (Dablanc, 2007; Woxenius, 2012); collaborative actions among the companies are hindered (Lindholm, 2013). The implementation of city
logistics measures (Thompson and Taniguchi, 2001), for which collaboration among partners is essential (Lindholm and Behrends, 2012),
has revealed that increasing accessibility in urban distribution centres
results in a decrease in the number of deliveries per tour (Nuzzolo and
Comi, 2014). Therefore, supply and demand for logistic services may
be more balanced. Anderson et al. (2005) nd the impact of weight
and time restrictions on distribution operations is negative with regard to increases in vehicle rounds, total distance travelled, and long
queue at the receiver's premises. Joint actions between receivers and
operators would reduce the impact of the xed delivery schedule and
minimize delivery difculties imposed by general regulations (Munuzuri et al., 2005). Thus, we argue the following:
H3. Regulation negatively moderates the positive relationship
between lack of collaboration and load/unload interfaces.

H4a. Detour mediates the negative effect of load/unload interfaces

on delivery performance.
H5. Load/unload interfaces are negatively related to delivery
performance.
2.5. The impact of detour on logistical performance
Woxenius (2012) notes that trafc congestion, geography and
topography challenges, unsafe routes, and other causes are reasons
for not choosing the shortest path; more time and transportation
resources are spent in delivery (Woudsma et al., 2008). Firms
agree delivery time lost costs money, and more efcient logistical
operational performance, as measured in terms of delivery time
and distance required, ultimately benets the customer (Figliozzi,
2010). Therefore, companies attempt to maintain their competitiveness by preferring contracting partners that have expertise in
delivering their goods securely and with high logistical efciency,
diversifying their deliveries through the use of different sizes of
vehicles, and providing additional logistic services (Roorda et al.,
2010). However, detour may determine the diversity and total
amount of freight vehicle activity required to deliver a given
quantity and diversity of goods, suggesting trafc congestion and
lack of road infrastructure could negatively inuence delivery time
and result in disruption of goods (Browne, 1993) in urban areas
(Cherrett et al., 2012). Furthermore, detour can make it difcult to
optimise routes due to narrow street restrictions and complex
delivery issues (e.g., vehicle types, diversity of products, high drop
rates, lot sizing and lading factors, use of large eets), resulting in
lower efciency of freight journeys in terms of logistical performance. Thus, we arrive at our nal hypothesis:
H6. Detour is negatively related to logistical performance.

2.4. Load/unload interfaces in megacities

2.6. Control variable in freight distribution model

One of the major problems with the city logistics concept

(Thompson and Taniguchi, 2001) is the lack of specic infrastructure
for load/unload interfaces such as parking, loading/unloading of
cargo vehicles (Munuzuri et al., 2005), and long unloading queues
(Stathopoulos et al., 2012). Researchers often view delivery activities
as a major contributor to congestion and trafc problems in urban
areas (Cherrett et al., 2012). Truck restrictions imply reduced delivery
times for truck operations, because trucks cannot adopt multi-stop
consolidated deliveries and rms are forced to use several light vehicles (Castro and Kuse, 2005). Consequently, many more parking,

To mitigate potential bias in the sample, we included rm size

(determined by number of employees) in the structural model as a
control variable affecting logistical performance (Cruijssen et al.,
2007). In our research, we posit that small operators active in logistics are likely to operate in market niches in which cooperation
with large rms is not possible or protable (Cruijssen et al.,
2007). In contrast, larger service ranges and customer bases provide large companies with more opportunities to develop their
activities more efciently. Many large transport operators contract
small sub-contractors to make their deliveries and collections

40

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

Fig. 1. Research framework.

(Selviaridis and Spring, 2007). The sub-contractors focus on restricted area or market-share, although they are coupled with
tendency for use of illegal load/unload and parking areas (Stathopoulos et al., 2012). We expect that small and medium operators operate more efciently in large urban centres than large
rms because they have more knowledge of the trafc and local
and unsafe areas and more rapport with retailers.

3. Methodology
The methodology of this study consists of a survey distributed
to freight operators (third-party logistics [3PLs] and carriers). The
3PLs are responsible for planning and managing the physical distribution of goods to shippers: they consolidate/deconsolidate,
transport and warehouse goods, provide inventory management
and information related to tracking and tracing, contract carriers
to transport the goods to retailers, and develop secondary activities such as assembly and installation of products (van Laarhoven
et al., 2000). Carriers are companies that provide inbound and
outbound transportation, door-to-door transportation service,
contracted delivery and customised services (such as delivery from
temporary storage), cross-docking systems, and distribution in
various small vehicles. These two types of companies, collectively
called freight operators, focus on distribution of goods in the
SPMR, which comprises 39 municipalities and is the sixth largest
metropolitan region in the world. So Paulo is the largest municipality in the SPMR and has approximately 12 million inhabitants.
The SPMR has approximately 20 million inhabitants and has increased in size since 2005 (De Vasconcellos, 2005); thus, the SPMR
effectively represents a high-density urban area.
The survey consists of two questions related to freight distribution constructs and logistical performance by the companies
inside the metropolitan region (see the Appendix). Developed
from our literature review, the survey addresses the follow research questions:
Which freight distribution constructs most affect company logistical performance?
Which constructs are the most relevant in freight distribution
systems?
To investigate the interactions among the constructs, based on
actors' perceptions, we propose a structural equation model (SEM)

that explains the interaction among them, and the difculties of

distributing goods in a metropolitan region from the company
perspective. Similarly, Deng. et al. (2013) apply an SEM to investigate the impact of three sets of port factors (port demand,
port supply, and value added activity in port) on the development
of a regional economy from a logistics perspective. The authors
relate widely used studies in the logistics area and also highlight
the use of SEM as an important analytical tool to investigate empirical research in the logistics area. Other relevant studies also
focus on transport system perceptions but do not take into account
the perspectives of the companies (Harvey et al., 2014). Therefore,
to support these questions and our SEM, the following subsections
describe the conceptual model and the survey research procedure.
3.1. Conceptual model: latent and indicator variables
Structural equations represent a set of linear equations, in which a
set of variables are related to each other in terms of cause and effect
(causal models) or paths through ordered networks of statistical dependence (path analysis). There are two types of SEM: covariancebased SEM, primarily used to conrm or reject theories based on
multiple variables empirically tested, and PLS-SEM, primarily used to
develop theories focusing on explaining the variance in the dependent
variables in the model (as in our case). The PLS-SEM consists of two
steps. First, a structural model (inner model) represents the constructs
(circle in our model, Fig.1) and displays the relationships among
constructs. Second, the measurement model (outer model) displays
the relationship between each construct and their indicator variables
(rectangles in our model Fig.1; Hair et al., 2014).
Therefore, based on our literature review, we detail below H1, H2,
H4, H5, H6 as the direct hypotheses, H1a, H1b, H2a, H2b, H4a as the
hypotheses regarding indirect and mediated relationships, and H3 as
the moderate hypothesis. Fig. 1 presents the model linking freight
distribution constructs and logistical performance constructs (grey
area) that we test empirically in this study; rm size is a control
variable. The Appendix presents all of the indicator variables for each
latent construct. In the model, we use reective latent constructs,
which can be reectively measured through their respective indicators. For example, collaboration (latent variable) is represented by
X3, X4 and X5 (indicators).

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

3.1.1. Measuring the effect of regulations

In our model, regulations are an important antecedent in detour (Castro and Kuse, 2003, 2005), which in turn inuences the
logistical performance of the companies (Woxenius, 2012). Moreover, we expect the moderating effect (H3) of regulations on the
relationship between collaboration and load/unload interfaces, as
detailed earlier.
3.1.2. Measuring the effect of collaboration
Collaboration between freight operators and receivers is an
important antecedent in load/unload interfaces (Stathopoulos
et al., 2012; Cherrett et al., 2012), which in turn inuence the logistical performance of the company (Castro and Kuse, 2005) and
detour (Cherrett et al., 2012).
3.1.3. Measurement of load/unload interfaces
The load/unload interfaces are antecedents in detour (Woudsma, et al., 2008). We also expect load/unload interfaces to have a
mediating effect on the relationship between regulations and detour and on the relationship between collaboration and detour,
because the delivery activities cause trafc problems (Cherrett
et al., 2012) and clashes of interest among logistic companies
(Munuzuri et al., 2005).
3.1.4. Measurement of detour
Detour is a relevant antecedent in logistical performance
(Roorda et al., 2010; Browne, 1993). We also expect detour to
present a mediating effect on the relationship between load/unload interfaces and logistical performance (Munuzuri et al., 2005).
3.1.5. Measurement of logistical performance indicators (LPI)
Logistical performance indicators (LPIs) are an efcient and
important measurement in freight distribution systems in that the
majority of a megacity's goods are served by road transport modes.
By studying LPIs, we can gain a better understanding of companies' metrics to optimise their logistics and transport activities
because LPIs focus on time and characteristics of delivery (Hesse
and Rodrigue, 2004). Time indicators can identify the performance
of the operators during deliveries in the supply chain (Castro and
Kuse, 2005). We investigate time indicators based on on-time-infull and fullment schedule delivery. Characteristics of delivery
(e.g., frequency, volume, type of product, distance) as they affect
logistical performance are also important in the megacity context.
For example, high-frequency deliveries within a stricter time
window demand a large eet of lighter vehicles as well as detailed
route planning. In the SPMR and other large cities in developing
countries, productsespecially electronic and vehicle spare parts
have been stolen as they are being delivered; consequently, escort
vehicles are needed to inhibit the thefts. Therefore, we consider
indicators regarding the availability of vehicles and freight theft
(Lu, 2003).
3.1.6. Firm size as a control variable
The construct of relationships involved in logistical performance is controlled by the size of the rm (Cruijssen et al., 2007).
We test whether small and medium rms operate more efciently
than large rms in a local area with a high number of urban issues,
due to taking collaborative action with retailers to deliver their
goods ahead of or behind schedule, and whether they tend to
develop expertise in local delivery related to security and access.
3.2. Data analysis
We t the model using a partial least squares (PLS) procedure
based on the software package SmartPLS 2.0 (Ringle et al., 2005),
which we used to estimate both the measurement and the

41

structural models. We chose PLS because the sample size was

relatively small (Chin, 1998, p. 295) and because some of the
measured variables showed highly uneven distributions. One
advantage of PLS in comparison to covariance-based approaches to
structural equation modelling (SEM) is that PLS does not presume
any distributional form for measured variables (Chin, 1998, p. 295).
Furthermore, PLS can be applied more easily to relatively small
data sets, as it requires only ten cases per predictor in the regression with the largest number of parameters (Chin, 1998,
p. 311).
3.3. Sample and procedures
Initially, we obtained a list of 6500 potential contacts from So
Paulo and the Regional Association of Cargo Carriers. We further
rened the list to a population of roughly 1100 carriers that deliver
goods within the SPMR and 43 3PLs, from September to December
2013. Thus, an innite population was considered and a convenience sampling was used. We received 138 carrier responses, of
which 95 were valid, and 30 TPL responses, of which 24 were
valid. We did not consider incomplete questionnaires and inadequate responses (i.e., the company worked with e-commerce
services or the respondent did not work in the logistical sector).
Most sample respondents ranked as senior managers, directors or
above (79.59%) and had an average 12 years of work experience. The
industrial sector was evenly distributed among the food and drink,
vehicle spare parts, telecommunication/energy, electronic, hygiene,
chemical and pharmaceutical categories, which comprised 84% of the
work for industrial respondents. Sizes of carriers and 3PLs, in terms of
number of employees and annual revenue measured by service activities, were as follows: 83.33% of the 3PLs were large (more than 100
employees and more than $1.2 million in annual revenue); 67.37% of
the carriers had fewer than 100 employees; and 55.79% generated
annual revenues greater than $1.2 million. Many of the carriers had
domestic capital (97.89%) and were located inside the megacity
(76.84%), whereas 50% of the 3PLs were foreign companies and 45.83%
worked inside the megacity.
To verify the statistical adequacy of the sample (i.e., to determine whether 119 questionnaires would be sufcient to analyse
the data), we used the software GnPower 3 (Faul et al., 2007).
According to the search parameters to be performed, as recommended by Cohen (1988) and Hair et al. (2014), we needed
power0.80 and f2 (Effect size) 0.15 (medium effect). The software indicated the minimum sample size required for the validity
of the research would be approximately 77 questionnaires. We
concluded the sample did not exhibit any problems, was of sufcient size, and represented rms of various industries that delivered goods in metropolitan areas.
3.4. Measures
We developed the questionnaire on the basis of Churchill's
questionnaire (1991) and from research questions and various instruments used by the participating companies to measure their
daily logistical performance.1 We used a Likert scale ranging from 1
(strongly disagree) to 7 (strongly agree) to measure the actors'
perceptions related to each item of the freight distribution and their
logistical performance constructs; a 2-point scale was used to convert rm size as control variables (see the Appendix). As Forza (2002)
advises, we asked experts to review the pre-test and nal questionnaire. We administered the pre-test to ve companies (DHL,2
1

The full version of the questionnaire is available upon request to the authors.
Transport manager of a third-party logistical (telecommunication, electronic,
hygiene, food and drink).
2

42

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

Table 1
Scale validity and reliability.
Scales Itemsa

Factorial loadings

Highest cross loading

Regulatory
Zone circulation for cargo vehicles
Restricted timetable

0.93
0.92

0.53
0.52

Collaboration
Lack of sharing information with retailers
Lack of exibility for retailer delivery in another time
Lack of commitment for receiving the goods

0.77
0.82
0.79

0.53
0.58
0.67

Load/unload interfaces
Long queue to load/unload
Lack of suitable loading/unloading areas
Lack of suitable parking area

0.82
0.87
0.85

0.62
0.63
0.67

Detour
Lack of security/risky areas
Narrow streets to delivery
Intense trafc/Congestion

0.77
0.85
0.82

0.59
0.56
0.62

Logistical performance
Freight theft/robbery occurs
Vehicle availability
On-time and in-full delivery
Fullment schedule delivery

0.63
0.89
0.64
0.84

0.02
 0.12
0.00
 0.07

a
b
c

4. Results and empirical analysis

4.1. Reliability and validity
We used conrmatory factor analysis (CFA) to simultaneously
validate the measures of all variables used in our research. We
checked the measurement properties of the constructs in the
model by assessing convergent validity, reliability, and discriminant validity.
To assess convergent validity, we examined factor loadings of
scale items on their corresponding constructs. The CFA indicated the
items loaded signicantly (40.5) and their values were high on their
respective constructs. We also determined the average variance extracted (AVE), composite reliability (CR), and Cronbach's alpha (see
Table 1). We assessed scale reliability using an alpha coefcient of
0.6 and a composite reliability (CR) index of 0.7 (Hair et al., 2014). As
Table 1 indicates, all CR estimates and alpha coefcients were above
their respective cut-offs. The item loadings and the overall model t
results suggest acceptable unidimensionality and convergent validity
Supply chain manager of distribution centre (spare parts of vehicles).
Director of third-party logistical (consumer package goods sector, specically
chemical, and chemist categories).
5
Director of distribution centre (consumer package goods, specically clothing
and accessories).
6
Director of transportation (consumer package goods, specically optical,
luxury and sports eyewear products).
4

CRc

Cronbach's alpha

0.85

0.92

0.83

0.63

0.83

0.70

0.72

0.89

0.81

0.66

0.85

0.74

0.59

0.84

0.77

Scale item abbreviation: same as in Appendix.

Variance extracted.
Composite reliability.

GM,3 Stralog,4 Daft5 and Luxottica6) and to some carrier partners

that provide exclusive service for the companies. We administered
the nal questionnaire to freight operators using the Survey Monkey
electronic platform to obtain their perceptions of freight distribution
issues and LPIs. To this end, we asked the informants if they delivered
goods in the SPMR and invited them to answer the questionnaire
with focus on that region.

AVEb

for the measures (Bollen, 1989).

We tested the discriminant validity of the research scales in
three ways that assessed AVE, correlation values, and item crossloadings. First, for each scale, we determined whether item crossloadings were lower than their factor loadings (Table 1). Second,
we conrmed all correlation values were less than 1 by an amount
greater than twice their respective standard error. Third, for each
scale, we determined whether the square root of the AVE was
higher than the individual correlations with other latent variables
(Fornelll and Larcker, 1981) (see Table 2). In addition, the correlations between the constructs did not present high values.
In addition, to ensure common method bias (CMB) did not affect the interpretation of results, we included a common method
factor in the PLS-PM model, that linked to all latent variables
(single-indicators) converted from observed indicators (Liang
et al., 2007). The results showed method factor loadings were not
signicant and the indicators' substantive variances were substantially greater than their method variances. In light of these
results, we found sufcient evidence for convergent and discriminant validity exists, and concluded CMB was unlikely to be a
serious concern.
4.2. SEM
Using the previously mentioned measurement model results,
we conducted an estimated model of constructs related to freight
distribution and logistical performance indicators from the freight
operators' viewpoints. Six evaluative dimensions (regulations,
collaboration, detour, load/unload interfaces, regulations as a
moderator, and rm size) acted as exogenous variables, and
logistical performance represented the endogenous variable.
Because the model uses reexive constructs, we used
Tenenhaus et al. (2005) goodness-of-t (GOF) measure to evaluate
the overall t of the model. This measure is dened as the geometric mean of the average communality, which measures the

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

Table 2
Discriminant validity for the model.

#1. Regulations
#2. Collaboration
#3. Load/unload interfaces
#4. Detour
#5. Logistical performance
a
b

#1

#2

#3

#4

#5

0.92a
0.39b
0.44
0.57
 0.13

0.79
0.75
0.65
 0.15

0.85
0.71
 0.09

0.81
 0.20

0.77

Numbers shown in bold on the diagonal denote the square root of the AVE.
The other numbers represent correlations between latent variables.

percentage of variance in a given original variable explained by its

latent variable, and the average of Square root (R2), for endogenous constructs, as follows:

GOF =

(Mean (Communality)x) Mean (R 2)

Table 3 shows the global GOF value of 0.56, which exceeds the
cut-off value of 0.36 for large effect sizes of R2 (where AVEZ 0.5),
as proposed by Wetzels et al. (2009), who argue the GOF criterion
proposed for PLS path modelling has a diagnostic purpose of the
global measurement of the model. Recently, researchers have also
used this measure as a global measurement of the model (Bloemer
et al., 2013; Latan and Ghozali, 2002). The Q2 (Stone-Geisser indicator), which measures the predictive relevance of the model
(i.e., the adjusted model accuracy), produces values greater than 0;
therefore, it performs suitably for this research (Hair et al., 2014).
The R-square values may be considered suitable for this research.
According to Chin (1998, p. 323), R2 0.67, 0.33 and 0.19 in PLS-PM
produces strong, moderate and modest effects, respectively. In behaviour science, the R-square can be classied as low effect (R2 2%),
medium effect (R2 13%) and large effect (R2 26%) (Cohen, 1988). In
accordance with Ringle et al. (2009), R-square moderate values are
acceptable when one latent variable is explained for one or two
exogenous variables. In addition, Hair et al. (2003) strongly suggests
the R-square should be analysed through its practical signicance. In
line with these classications and suggestions, we conclude the Rsquare in Table 3 is suitable for this model.
Moreover, we found the variance ination factor test value for
all variables was lower than 3.3 (Diamantopoulos and Siguaw,
2006), which suggests no multicollinearity exists in the model
(Lowry and Gaskin, 2014). Note that we assessed the multicollinearity before testing for mediation regulation, collaboration,
and load/unload interfaces variables. We measured the structural
model properties using beta coefcients, standard deviation error,
and t-values computed using 1000 bootstrapping runs. Fig. 2
shows all path analyses.
The results support the hypothesized effects of the constructs.
We highlight the large effect of collaboration on load/unload interfaces (H2 0.58): lack of collaboration between freight operators and retailers in terms of information sharing, exibility, and
commitment in receiving goods contributes to queue increases,
limited parking, and use of unsuitable areas for loading/unloading
Table 3
R2, Communalities and GOF values for the model
Latent variables

R2

Communalities GOF Predictive relevance (Q2)

Regulations
Collaboration
Load/unload interfaces
Detour
Logistical performance
Firm size
Average

0
0
0.64
0.61
0.16
0
0.47

0.83
0.63
0.72
0.66
0.59
1
0.66

0.86
0.62
0.45
0.39
0.09

0.56d

43

used by trucks or smaller vehicles in the retailer area. Lack of

collaboration also contributes to the necessity of maintaining a
dedicated team to provide extra service. Indirectly, the construct
collaboration exerts more impact on detour (H2a 0.24); in other
words, lack of collaboration increases trafc congestion and the
use of narrow streets and unsafe areas.
To test the mediating effect in our path model, we followed the
simple test of mediation proposed by Baron and Kenny (1986). The
results show detour is a full mediator between relationship regulation and logistical performance. In this case, the full mediation
occurs when the regulation no longer has a signicant effect on
the logistical performance when the mediator, detour, is included
in the model. Therefore, this result provides strong evidence that
detour predicting only a direct relationship between regulations
and logistical performance is suboptimal and theoretically incorrect. Load/unload interfaces is a full mediator between relationship collaboration and logistical performance, when the
same procedure is applied to collaboration as predictor, criterion
variable, logistical performance, and intervening variable load/
unload interfaces. The results indicate load/unload interfaces
partially mediate the effect of collaboration and regulations on
detour. Partial mediation occurs when the predictor has a signicant effect, but its effect is diminished when the mediator is
included in the model (Baron and Kenny, 1986). Our observations
show only indirect effect exists between load/unload interfaces
and logistical performance by detour.
In our model, the construct load/unload interfaces contribute
most to increasing the total effect (direct and indirect effect),
whereas collaboration stresses the direct effect. Collaboration,
regulation, and load/unload interfaces contribute to less efcient
logistical performance, mainly when detour mediates this relationship. To investigate which construct causes most effect in the
model, we analysed the explained variance for each construct (see
Table 4). We found regulations, collaboration, and load/unload
interfaces explained 61.02% of the variance in detour. However, the
load/unload interfaces construct was solely responsible for 47.87%
of the total variance. In other words, our results show detour is
caused mainly by the lack of suitable parking and loading/unloading areas. In addition, collaboration emerges as the most important construct contributing to increase of load/unload interfaces (variance total 67.94%). Therefore, our results indicate little
joint action between partners and lack of exibility, commitment,
and information sharing.
The observed R2 shows the determining power of regulations in
detour can be considered large but low in load/unload interfaces;
the determining power of collaboration in detour is high and very
high in load/unload interfaces; the determining power of load/
unload interfaces in detour is lhigher; and the determining power
of detour in logistical performance is lower. Moreover, larger rm
size produces a negative impact in logistical performance
(  0.335; t-value 4.13**, f2 0.13). In other words, small rms
have better logistical performance than large ones.
Moreover, we examined the moderating effect of regulations
(negative effect) on the positive relationship between collaboration and load/unload interfaces by multiplying each moderator
variable with each predictor variable to add the interaction terms
to the main effects model. The results provide support for the high
moderating effect of regulations (H3; f2 0.19). Table 5 summarizes the results of the hypotheses testing.

5. Discussion of results
With regard to individual constructs, we note detour has the
most impact on company logistical performance. However, this
construct becomes more complicated when load/unload interfaces

44

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

Fig. 2. Freight distribution research model.

Table 4
Explained variance for each construct of the structural model.
Relationships

% VAR % Total
VAR

Effect size
(f2)n

Regulations - Detour (H1)

Collaboration - Detour
Load/unload interfaces - Detour (H4)

17.01
14.80
29.21
61.02

27.88
24.25
47.87
100.00

0.21
0.17
0.41

Collaboration - Load/unload interfaces (H2) 43.24

Regulations -Load/unload interfaces
4.61
Collaboration*Regulations-Load/unload in- 15.79
terfaces (H3)
47.85

67.94
7.25
24.82

0.76
0.05
0.19

Detour - Logistical performance (H7)

Firm Size - Logistical performance (control
variable)

100.00

5.08
11.50

30.65
69.35

16.58

100.00

0.05
0.13

n
Effect size contribution of each constructs separately, values according to
Cohen (1988).

problems are present, i.e., lack of suitable areas to park and load
and unload freight. The importance of this construct indicates that
although increasing accessibility reduces the number of deliveries
per tour, it also increases the probability of more round trips for
light goods vehicles and small delivered quantities (Nuzzolo and
Comi, 2014).
According to the effects shown by our model, regulation, along
with lack of collaboration, appear to be the Achilles' heel of freight
distributors, in that both factors contribute (directly and

indirectly) to detour, which results in less efcient logistical performance. These are the most relevant factors in the freight distribution system; addressing them could provide harmonisation
between government and business. Governmental regulations are
typically restricted to detour and lack of collaboration. However, to
a lesser extent, perceptions of load/unload interfaces involving
lack of collaboration follow when restrictive regulations (as a
moderator) are in place. Although the regulations reduce the negative effect of collaboration in load/unload interfaces, other factors may contribute to lack of parking and loading/unloading
areas, such as (1) more vehicles in circulation, which necessitates
more space; (2) imbalance between demand for infrastructure to
park/load/unload and obsolete regulatory policy (Dablanc, 2007);
(3) lack of locally specic regulations (Munuzuri et al., 2005); or
(4) lack of systemic vision in planning the freight distribution
system (Crainic et al., 2004). Our model shows small/medium
companies provide the best logistical performance in megacities,
possibly because they have more knowledge of how to full customer requirements and focus on distribution to specic markets
than larger companies located outside the megacity.
Our model indicates detour is directly and indirectly caused by
governmental regulations and indirectly caused by lack of collaboration between partners. In turn, detour contributes to a decrease in rms' logistical performances in terms of time spent on
deliveries, theft due to use of unsafe areas, and difculty in advance route planning. It also hinders eet planning and availability
of vehicles if frequent and restricted time deliveries in remote and
unsafe areas are demanded. Therefore, the interaction of the
freight distribution model should be viewed as a whole system.
We cannot focus on one construct without taking others into account. For example, although detour is an outcome of trafc

Table 5
Summary of the results of hypotheses testing.
Hypothesized directions

Effects

Affected

Results

H1. Regulations
H1a. Regulations - Load/unload interfaces
H1b. Regulations - Detour
H2. Collaboration
H2a. Collaboration- Load/unload interfaces
H2b. Collaboration- Load/unload interfaces
H3. Collaboration*Regulations
H4. Load/unload interfaces
H4a. Load/unload interfaces - Detour
H5. Load/unload interfaces
H6. Detour

Direct
Partially mediating
Full mediating
Direct
Partially mediating
Full mediating
Moderating
Direct
Indirect
Direct
Direct

Detour
Detour
Logistical performance
Load/unload interfaces
Detour
Logistical Performance
Load/unload interfaces
Detour
Logistical performance
Logistical performance
Logistical performance

Supported
Supported
Supported
Supported
Supported
Not supported
Supported
Supported
Supported
Not supported
Supported

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

congestion, and physical obstacles such as narrow streets and

unsafe areas to deliver goods, it appears to be more important
when it is connected with other constructsfor example, regulations, rms' logistical performances, and, indirectly, collaboration.
By viewing the problem of congestion bottlenecks too narrowly,
without considering the system as a whole, we may nd only
partial solutions that address only trafc congestion and not the
entire construct (e.g., cargo theft, impact of the detour on logistical
performance). Crainic et al. (2004) argue the authorities have not
paid attention to the transportation of goods in the city as a whole,
except through parking, street access, and restriction of hours of
commercial truck operation. Others investigations could also be
treated in the model. Congestion charging, and actions based on
driver behaviour theory (Holgun-Veras, 2006), have been successful in reducing congestion (Holgun-Veras, 2010).

6. Conclusions
Our work innovates by providing a theoretical understanding of
the complex freight distribution system, taking into account the
main stakeholders, governmental regulations, collaboration between actors, and issues and logistical performances of the companies. Rather than focusing on the regulations or collaboration as
separate freight distribution antecedents, our model provides insights into how the positive effect of collaboration on load/unload
interfaces can be moderated by regulations. The ndings provide
theoretical understanding of how freight distribution affects logistical performance of large companies. In this case, the results
reveal regulations and lack of collaboration appear to be the most
important constructs in the model; they show detour has the most
signicant impact on rm performance.
In terms of collaboration, companies could improve their performance by investing in information technology systems and
increasing their joint actions to align the interests of the various
companies involved in the distribution system. For example,
freight operators could strengthen relations with their shipper
partners to get information about retailers and receivers regarding
local parking, accessibility, regulations, and exibility to receive
the goods in off-time windows. They could then plan their logistic
activities in terms of load and availability of vehicles, contracting
of assistants, and delivery type (single-versus multi-stop deliveries). Information systems involve vendor-managed systems,
technology to track goods and vehicles, and route planning. Firms
could also invest in conducting periodic meetings to share demand
forecasting, improve eet and route planning, and solve some
problems related to factors such as delay, mistakes, damaged
goods, and theft. Moreover, technical visits would be an excellent
means for partners to develop knowledge of each other's logistic
reality and to make better operational and strategic decisions
(Vieira et al., 2009). Along these lines, receivers and freight operators could check suitability of delivery bays for different vehicle
sizes, local trafc idiosyncrasies, necessity of staff to receive goods,
exibility to receive goods off schedule, and intermediate rental
places (with sharing costs) to temporarily house stock or take the
place of the retailer Zelst et al. (2009).
From a governmental point of view, separation of restricted
timetables and zone circulations for cargo does not work well; this
approach needs to be updated. The message of the model is that
all constructs must be analysed together. Our research provides
support for government initiatives to implement new proceedings
or rules to alleviate issues related to intense trafc, vehicular accidents, disorganised land use, and cargo thefts, and shows where
the investment in urban infrastructure take place, in terms of
mobility and freight distribution. Our model identies issues that
have a great impact on companies located in the SPMR, but it is

45

also relevant to cities around the world.

Our research contributes theoretically to the relevant literature
on freight distribution by including policy measures (such as
regulations), private initiatives (such as collaboration), and (to a
limited extent) safety aspects related to detour and logistical
performance, in a single analysis. Our model jointly analyses a
number of relevant constructs. Moreover, from the collaboration
point of view, this research sheds light on aspects related to
commitment, information sharing, and exibility that have been
little studied in the freight distribution research eld (Lindholm,
2013).
6.1. Limitations and further research
Although our research provides suitable statistical results, we
must clarify some limitations, related mostly to the sample and
data collection. First, the sample of our proposal is restricted to the
SPMR, although the results may be useful for other large cities in
the world. The developed model might be useful to governments
to address policies and to companies to invest in collaborative
actions. Nevertheless, in a detailed analysis the model should
consider other stakeholder perspectives such as those of retailers
and shippers. The model should also take into account more operational variables such as time windows to delivery, necessity of
overnight delivery, delivery lead time, lighter vehicle types, diversity of products, high drop rates, lot sizing, and lading factors.
Second, related to the concerns of Bloemer et al. (2013), we used
only one questionnaire to measure all relevant constructs. Therefore,
although the single-method common bias does not appear to be an
issue for our data, it is not completely certain that some relationships
between constructs are over-or-underestimated due to affected error
distributions. Testing for the reliability and the validity of the constructs, especially discriminant validity, also serves to diminish the risk
of incorrect estimation of the relationships between the constructs;
nonetheless, more research would strengthen our ndings.
Third, the respondents provided self-assessments, which results in two problems. The rst is the responses represent viewpoints of the people who work in the companies. To minimize this
issue, we instructed respondents to ll out the questionnaire from
the perspective of the organisation. Moreover, to decrease the risk
of self-assessment bias, we ensured all scales for measuring the
different constructs were previously validated by managers and
were previously used in literature. However, further research
could use these documents to address and also aggregate other
variables to investigate other megacities. The second problem is
self-assessments affect how logistical performance indicators are
measured. To mitigate this factor, we used the words rarely and
always with each observed factor to constrain respondents to an
organizational viewpoint (e.g., Freight loss occurs rarely).
Fourth, the responses were collected on a web platform, with
many freight operators (mainly carriers) that serve other regions
outside SPMR lling out the questionnaire. To address this shortcoming, we asked informants whether they delivered goods to the
SPMR and invited them to answer the questionnaire focusing on
those deliveries. This procedure resulted in a roughly 30% refusal
rate and 60 call-backs.

Acknowledgements
This work was supported by the Brazilian agencies: CAPES
(Grant # 2480-13-1) and CNPq (Grant # 301140/2013-8).
This research was conducted while Dr Vidal Vieira was a
visiting postdoctoral researcher at the School of Industrial
Engineering at Eindhoven University of Technology.

46

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

Appendix
See appendix Table 6.

Table 6
Attributes and basic statistical results.
1) Research question: The follow items are related to freight distribution in
SPMR. Please indicate your agreement level in each one: 1 Strongly disagree; 2 Disagree; 3 Somewhat disagree; 4 Neither agree nor disagree; 5 Somewhat agree; 6 Agree; 7 Strongly agree.
Freight distribution constructs/
Indicator variables
Regulatory
X1. Zone circulation for cargo
vehicles
X2. Restricted timetable

Collaboration
X3. Lack of sharing information
with retailers
X4. Lack of exibility for retailer
delivery in another time
X5. Lack of commitment for receiving the goods

Load/unload interfaces
X6. Long queue to load/unload
X7. Lack of suitable loading/unloading areas
X8. Lack of suitable parking area

Detour
X9. Lack of security/risky areas
X10. Narrow streets to delivery
X11 Intense trafc/Congestion

Mean value

Standard deviation

5.82

1.86

5.59

2.01

4.94

1.69

5.70

1.45

5.50

1.64

5.92
5.98

1.47
1.47

5.92

1.59

5.94
5.29
6.47

1.60
1.63
1.19

2) Research question: Please indicate below your agreement level to your logistical performance in SPMR: 1 Strongly disagree; 2 Disagree; 3
Somewhat disagree; 4 Neither agree nor disagree; 5 Somewhat agree; 6
Agree; 7 Strongly agree
Logistical performance
indicators

Mean value

Standard deviation

X12.
X13.
X14.
X15.

4.50
4.26
4.83
5.24

1.84
1.87
1.61
2.09

Sample

Frequency (%)

68
51

57
43

Freight theft/robbery occurs

Vehicle availability
On-time and in-full delivery
Fullment schedule delivery

3) Control variable: Firm size

Number of employee
Small/medium rms
Large rms

a
Based on (Sebrae, 2013) classication which establish the size of the rm,
focusing in services, in number of employees into four groups: micro (119
employees), small (2099 employees), medium (100499 employees) and large
(500 or more employees). We aggregate micro, small and medium in one group
and large rm in another one.

References
Anderson, S., Allen, J., Browne, M., 2005. Urban logisticshow can it meet policy
makers' sustainability objectives? J. Transp. Geogr. 13 (1), 7181. http://dx.doi.
org/10.1016/j.jtrangeo.2004.11.002.
Baron, R.M., Kenny, D. a, 1986. The moderator-mediator variable distinction in

social psychological research: conceptual, strategic, and statistical considerations. J. Personal. Soc. Psychol. 51 (6), 11731182.
Behrends, S., Lindholm, M., Woxenius, J., 2008. The impact of urban freight transport: a denition of sustainability from an actors perspective. Transp. Plan.
Technol. 31 (6), 693713. http://dx.doi.org/10.1080/03081060802493247.
Bloemer, J., Pluymaekers, M., Odekerken, A., 2013. Trust and affective commitment
as energizing forces for export performance. Int. Bus. Rev. 22 (2), 363380.
http://dx.doi.org/10.1016/j.ibusrev.2012.05.002.
Bollen, K.A., 1989. Structural Equations with Latent Variables. Wiley, New York, p.
514.
Browne, M., 1993. Logistics strategies in the single European market and their
spatial consequences. J. Transp. Geogr. 1 (2), 7585. http://dx.doi.org/10.1016/
0966-6923(93)90001-G.
Castro, J.T., Kuse, H., 2003. A study on the impact and effectiveness of the truck ban
scheme in metro Manila. J. East. Asia Soc. Transp. Stud. 5, 21772192.
Castro, J.T., Kuse, H., 2005. J. East. Asia Soc. Transp. Stud. 6, 29472962.
Cherrett, T., Allen, J., McLeod, F., Maynard, S., Hickford, A., Browne, M., 2012. Understanding urban freight activity-key issues for freight planning. J. Transp.
Geogr. 24, 2232. http://dx.doi.org/10.1016/j.jtrangeo.2012.05.008.
Chin, W.W., 1998. The partial least squares approach to structural equation modeling. In: Marcoulides, G.A. (Ed.), Modern Business Research Methods. Lawrence
Erlbaum Associates, Mahwah, New Jersey, pp. 295336.
Churchill, G.A., 1991. Marketing Research: Methodological Foundation, 5th edition.
Dryden Press, New York.
Cohen, J., 1988. Statistical power analysis for the behavioral sciences, The Concept
of Power Analysis., second ed Psychology Press, New York
Cidell, J., 2010. Concentration and decentralization: the new geography of freight
distribution in US metropolitan areas. J. Transp. Geogr. 18 (3), 363371. http:
//dx.doi.org/10.1016/j.jtrangeo.2009.06.017.
Crainic, T.G., Ricciardi, N., Storchi, G., 2004. Advanced freight transportation systems
for congested urban areas. Transp. Res. Part C 12, 119137. http://dx.doi.org/
10.1016/j.trc.2004.07.002.
Cruijssen, F., Cools, M., Dullaert, W., 2007. Horizontal cooperation in logistics: opportunities and impediments. Transp. Res. Part E: Logist. Transp. Rev. 43 (2),
129142. http://dx.doi.org/10.1016/j.tre.2005.09.007.
Dablanc, L., 2007. Goods transport in large European cities: Difcult to organize,
difcult to modernize. Transp. Res. Part A: Policy Pract. 41 (3), 280285. http:
//dx.doi.org/10.1016/j.tra.2006.05.005.
Deng., P., Lu, S., Xiao, H., 2013. Evaluation of the relevance measure between ports
and regional economy using structural equation modeling. Transp. Policy 27,
123133. http://dx.doi.org/10.1016/j.tranpol.2013.01.008.
De Vasconcellos, E.A., 2005. Urban change, mobility and transport in So Paulo:
three decades, three cities. Transp. Policy 12, 91104. http://dx.doi.org/10.1016/
j.tranpol.2004.12.001.
Diamantopoulos, A., Siguaw, J., 2006. Formative versus reective indicators in organizational measure development: a comparison and empirical illustration. Br.
J. Manag. 17 (4), 263282. http://dx.doi.org/10.1111/j.1467-8551.2006.00500.x.
Domnguez, A., Holgun-Veras, J., Ibeas, ., dellOlio, L., 2012. Receivers' response to
new urban freight policies. Procedia Soc. Behav. Sci. 54, 886896. http://dx.
doi.org/10.1016/j.sbspro.2012.09.804.
Faul, F., Erdfelder, E., Lang, A.-G., Buchner, A., 2007. G*Power 3: a exible statistical
power analysis program for the social, behavioral, and biomedical sciences.
Behav. Res. Methods 39, 175191.
Figliozzi, M.A., 2010. The impacts of congestion on commercial vehicle tour characteristics and costs. Transp. Res. Part E: Logist. Transp. Rev. 46 (4), 496506.
http://dx.doi.org/10.1016/j.tre.2009.04.005.
Fornelll, C., Larcker, D.F., 1981. Evaluating structural equation models with unobervable variables and measument error. J. Mark. Res. 18 (1), 3950.
Hair, J.F.J., Hilt, T.M., Ringle, C.M., Sarstedt, M., 2014. A Primer on Partial Least
Squares Structural Equation Modeling (PLS-SEM). SAGE, Los Angeles, p. 328.
Hair Jr., J.F., Barbin, B., Money, A.H., Samouel, P., 2003. Essentials of Business Research Methods. Johns Wiley & Sons, Inc., United States of America.
Harvey, J., Thorpe, N., Caygill, M., Namdeo, A., 2014. Public attitudes to and perceptions of high speed rail in the UK. Transp. Policy 36, 7078. http://dx.doi.
org/10.1016/j.tranpol.2014.07.008.
Hesse, M., Rodrigue, J.P., 2004. The transport geography of logistics and freight
distribution. J. Transp. Geogr. 12 (3), 171184. http://dx.doi.org/10.1016/j.
jtrangeo.2003.12.004.
Holgun-Veras, J., 2010. The truth, the myths and the possible in freight road pricing
in congested urban areas. Procedia Soc. Behav. Sci. 2 (3), 63666377. http:
//dx.doi.org/10.1016/j.sbspro.2010.04.045.
Holguin-veras, J., Polimeni, J., Cruz, B., Xu, N., List, G., Nordstrom, J., Haddock, J.,
2005. Off-peak freight deliveries : challenges and stakeholders perceptions.
Transp. Res. Rec. 1906 (212), 4248.
Holguin-Veras, J., Wang, Q., Xu, N., Ozbay, K., Cetin, M., Polimeni, J., 2006. The
impacts of time of day pricing on the behavior of freight carriers in a congested
urban area: Implications to road pricing. Transportation Research Part A: Policy
and Practice 40 (9), 744766. http://dx.doi.org/10.1016/j.tra.2005.12.011.
Laarhoven, P. van, Berglund, M., Peters, M., 2000. Third-party logistics in Europe
ve years later. Int. J. Phys. Distrib. Logist. Manag. 30 (5), 425442. http://dx.
doi.org/10.1108/09600030010336216.
Latan, H., Ghozali, I., 2002. Partial Least Squares: Concepts, Techniques and Applications using SmartPLS 2.0 M3. Diponegoro University Press.
Liang, H., Saraf, N., Hu, Q., Xue, Y., 2007. Research article assimilation of enterprise
systems : the effect of institutional pressures and the mediating. MIS Q. 31 (1),
5987.

J.G. Vidal Vieira, J.C. Fransoo / Transport Policy 44 (2015) 3747

Lindholm, M., 2013. Urban freight transport from a local authority perspectivea
literature review. Eur. Transp./Trasp. Eur. 54, 137.
Lindholm, M., Behrends, S., 2012. Challenges in urban freight transport planninga
review in the Baltic Sea Region. J. Transp. Geogr. 22, 129136. http://dx.doi.org/
10.1016/j.jtrangeo.2012.01.001.
Lowry, P.B., Gaskin, J., 2014. Partial Least Squares (PLS) Structural Equation Modeling (SEM) for building and testing behavioral causal theory : when to choose
it and how to use it. IEEE Trans. Prof. Commun. 57 (2), 123146.
Lu, C.-S., 2003. The impact of carrier service attributes on shippercarrier partnering relationships: a shipper's perspective. Transp. Res. Part E: Logist. Transp.
Rev. 39 (5), 399415. http://dx.doi.org/10.1016/S1366-5545(03)00015-2.
Moberg, C.R., Cutler, B.D., Gross, A., Speh, T.W., 2002. Identifying antecedents of
information exchange within supply chains. Int. J. Phys. Distrib. Logist. Manag.
32 (9), 755770. http://dx.doi.org/10.1108/09600030210452431.
Morgan, R.M., Hunt, S.D., 1994. The commitment-trust theory of relationship
marketing. J. Mark. 58 (3), 2038.
Munuzuri, J., Larraneta, J., Onieva, L., Cortes, P., 2005. Solutions applicable by local
administrations for urban logistics improvement. Cities 22 (1), 1528. http://dx.
doi.org/10.1016/j.cities.2004.10.003.
Nuzzolo, A., Comi, A., 2014. Urban freight demand forecasting: aa mixed quantity/
delivery/vehicle-based model. Transp. Res. Part E: Logist. Transp. Rev. 65,
8498. http://dx.doi.org/10.1016/j.tre.2013.12.014.
Puckett, S.M., Hensher, D. a, 2009. Revealing the extent of process heterogeneity in
choice analysis: an empirical assessment. Transp. Res. Part A: Policy Prac. 43 (2),
117126. http://dx.doi.org/10.1016/j.tra.2008.07.003.
Ringle, C.M., Sinkovics, R.R., 2009. The use of partial least squares path modeling in
international marketing. Adv. Int. Mark. 20 (2009), 277319. http://dx.doi.org/
10.1108/S1474-7979(2009)0000020014.
Ringle, C.M., Wende, S., Will, A., 2005. SmartPLS (Version 2.0 - beta). University of
Hamburg, Hamburg.
Roorda, M.J., Cavalcante, R., McCabe, S., Kwan, H., 2010. A conceptual framework for
agent-based modelling of logistics services. Transp. Res. Part E: Logist. Transp.
Rev. 46 (1), 1831. http://dx.doi.org/10.1016/j.tre.2009.06.002.
Sebrae, 2013. Sebrae. Critrios de classicao de empresas: EIMEEPP. From

47

http://www.sebrae-sc.com.br/leis/default.asp?vcdtexto4154 (Retrieved
12.08.13).
Selviaridis, K., Spring, M., 2007. Third party logistics: a literature review and research agenda. The International Journal of Logistics Management 18 (1),
125150. http://dx.doi.org/10.1108/09574090710748207.
Stathopoulos, A., Valeri, E., Marcucci, E., 2012. Stakeholder reactions to urban
freight policy innovation. J. Transp. Geogr. 22, 3445. http://dx.doi.org/10.1016/
j.jtrangeo.2011.11.017.
Tenenhaus, M., Vinzi, V.E., Chatelin, Y.-M., Lauro, C., 2005. PLS path modeling.
Comput. Stat. Data Anal. 48 (1), 159205. http://dx.doi.org/10.1016/j.
csda.2004.03.005.
Thompson, R.G., Taniguchi, E., 2001. City logistics and transportation. Handbook of
Logistics and Supply-Chain Management. Elsevier, Amsterdam, pp. 393405.
Vieira, J., Yoshizaki, H., Ho, L., 2009. Collaboration intensity in the Brazilian supermarket retail chain. Supply Chain Manag. (1), 1121. http://dx.doi.org/10.1108/
13598540910927269.
Wetzels, M., Odekerken-Schrder, G., Van Oppen, C., 2009. Using pls path modeling
for assessing hierarchical construct models: guidelines and empirical illustration. MIS Q. 33 (1), 177195.
Woudsma, C., 2001. Understanding the movement of goods, not people: issues,
evidence and potential. Urban Stud. 38 (13), 24392455. http://dx.doi.org/
10.1080/00420980120094605.
Woudsma, C., Jensen, J.F., Kanaroglou, P., Maoh, H., 2008. Logistics land use and the
city: a spatial-temporal modeling approach. Transp. Res. Part E: Logist. Transp.
Rev. 44 (2), 277297. http://dx.doi.org/10.1016/j.tre.2007.07.006.
Woxenius, J., 2012. Directness as a key performance indicator for freight transport
chains. Res. Transp. Econ. 36 (1), 6372. http://dx.doi.org/10.1016/j.
retrec.2012.03.007.
Van Zelst, S., van Donselaar, K., Van Woensel, T., Broekmeulen, R., Fransoo, J., 2009.
Logistics drivers for shelf stacking in grocery retail stores : Potential for efciency improvement. Int. J. Prod. Econ. 121 (2), 620632. http://dx.doi.org/
10.1016/j.ijpe.2006.06.010.