Transportation Research Part A xxx (2016) xxxxxx

Contents lists available at ScienceDirect

Transportation Research Part A

journal homepage: www.elsevier.com/locate/tra

Role and potential of a trusted vendor certification program to

foster adoption of unassisted off-hour deliveries
Jos Holgun-Veras a,, Shama Campbell b, Lokesh Kalahasthi b, Cara Wang c
a
Rensselaer Polytechnic Institute, 110 Eighth St. JEC 4030, Troy, NY 12180, USA
b
Rensselaer Polytechnic Institute, 110 Eighth St. JEC 4027, Troy, NY 12180, USA
c
Rensselaer Polytechnic Institute, 110 Eighth St. JEC 4032, Troy, NY 12180, USA

a r t i c l e i n f o a b s t r a c t

Article history: The paper describes the research conducted by the authors regarding the willingness of
Available online xxxx receivers of supplies to accept unassisted off-hour deliveries (UOHD)deliveries made dur-
ing the off-hours without staff from the receiving establishment present at the time of
Keywords: deliveryfrom vendors certified by a Trusted Vendor Certification Program (TVCP). The
Off-hour deliveries main intent of the TVCP is to certify vendors as worthy of trust to conduct UOHD in a
Freight behavior research way that is both community friendly and that does not put the receiving establishment
Trusted vendor
at risk. To this effect, the authors collected attitudinal data from a sample of 450 receivers
Freight demand management
from two different metropolitan areas, who were asked whether or not they would be
inclined to accept UOHD if the vendors are certified as trusted vendor by a TVCP. The
paper conducts descriptive analyses of the data and estimates behavioral models to gain
insight into the industry sectors and establishment types that are the most inclined to
accept UOHD if a TVCP program is created.
2016 Elsevier Ltd. All rights reserved.

1. Introduction

An efficient urban freight system is essential to the well-functioning of modern cities, the quality of life of the citizens,
and the economy that support them. Notwithstanding its important role, the freight traffic required to transport the supplies
needed by both citizens and the rest of the urban economic activities produce large amounts of negative externalities, such
as congestion, pollution, traffic accidents, noise, and infrastructure damage. Conducting freight operations in a manner that
maximizes economic productivity and efficiency, and minimizes negative externalities, is crucial to enhance quality of life,
and achieve sustainability and economic goals. Freight Demand Management (FDM) seeks to achieve these goals by inducing
changes in the demand generated by the receivers of supplies. The emphasis on the receiversat the core of FDMis the
result of the considerable amount of freight behavior research that has been conducted in recent years (Holgun-Veras
et al., 2007, 2008; Holgun-Veras, 2011), though its roots are found in the theories of Spatial Price Equilibrium
(Samuelson, 1952) that explicitly consider the interactions between shippers, carriers, and receivers. The focus on receivers
is a natural progression of the approaches that started with the birth of City Logistics (Taniguchi et al., 1999), where the main
focus is on changing the behavior of the carriers.
The fundamental tenet of this paper is that enacting change on supply chains require the use of approaches that induce
the various agents involved to change behavior in the manner desired by policy makers. This perspective recognizes that the

Corresponding author.
E-mail addresses: jhv@rpi.edu (J. Holgun-Veras), campbs4@rpi.edu (S. Campbell), kalahl@rpi.edu (L. Kalahasthi), wangx18@rpi.edu (C. Wang).

http://dx.doi.org/10.1016/j.tra.2016.09.011
0965-8564/ 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
2 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

various agents in a supply chain have control over different decisions. A carrier that decides to change the color of its trucks,
or the type of engine they use, could do so without raising objections from shippers or receivers. As long as these changes are
within regulations, shippers and receivers are not likely to object because they are not materially impacted. In contrast, the
same carrier cannot unilaterally change the pick-up time or the delivery time because the concurrence of shippers and recei-
vers is required. After all, an uncooperative carrier may be easily replaced by a more accommodating one. The research con-
ducted has conclusively established the important role of receivers in determining the frequency, shipment size, and timing
of delivery. FDM seeks to employ different strategies to induce receivers to modify their freight demands in a socially ben-
eficial way.
One such strategy is the use of Off-Hour Delivery (OHD) programs to increase the amount of deliveries that takes place
between 19:00 and 6:00, in a way that does not lead to negative impacts on the local communities. The research conducted
for the implementation of OHD in the New York City (NYC) has shown that fuel consumption and emissions are significantly
lower during the off-hours compared to regular hours. The emission reductions associated with OHD are tremendous. Using
GPS data collected from trucks conducting regular-hour deliveries and OHD, Holgun-Veras et al. (in press-a) estimated that
emission reductions from OHD range between 55% and 67% depending on the pollutant. If 30% of the midsize trucks making
deliveries in the NYC metropolitan area shift to the off-hours, the ensuing emission reductions would amount to 8.66 million
tons/year of CO2, and 24.5 thousand tons/year of CO (Holgun-Veras et al., in press-a).
The chief insight is that OHD is relatively easy to implement if the agents involvedmost notably carriers and
receiversbenefit from it. It has been shown that carriers are naturally inclined to do OHD as they directly benefit from time
and monetary savings, do not have to contend with problems associated with finding suitable parking in a congested urban
area (Holgun-Veras et al., 2008), and are able to complete their deliveries without the added stress of competing with the
hustle and bustle that occurs during the day hours (Brom et al., 2011). However, for receivers, switching to OHD could lead to
additional costs (e.g., paying staff to be present, installing security devices) and/or added risk of property damage or theft, or
upsetting the residents of surrounding communities with noise and other night disturbances. To compensate the receiver for
these effects, both monetary and non-monetary incentives ought to be offered. Monetary incentives may include one-time
financial incentive such as a tax rebate or a cash payment, and vendor discounts that pass a portion of the savings from OHD
to the receiver. Non-monetary incentives may include public recognition of the business for participating in OHD, and/or
providing business support services to the receivers that accept OHD (Holgun-Veras et al., in press-b).
To identify real-life issues with OHD and to test the effectiveness of financial incentives to receivers, a pilot test was con-
ducted in NYC in 2009. About 35 companies participated in the test, which was deemed a success by all participants. About
half the receivers elected to use staffed OHD, i.e., staff of the receiving establishment is present at the time of delivery, and
the other half decided to use unassisted OHD, or UOHD, where the vendors are granted access to the establishment to make
OHD without supervision from the receiver. These two alternatives exhibit different tradeoffs between risk and reward. The
staffed alternative, using some form of human intervention to do OHD, minimize risk though at a cost; the UOHD minimize
costs though taking a measure of risk. Consistent with the prediction of economic theory, the receivers using staffed OHD
reverted back to regular hour deliveries upon termination of the financial incentive. In contrast, all but one of the receivers
using UOHD decided to continue receiving UOHD after the incentive ended. The main reason for this decision was the reli-
ability of UOHD, which allow receivers to reduce their inventory levels (and cost) and use their staff in a more efficient man-
ner. In addition, the pilot test experience led these receivers to recalibrate their perceptions of the risk associated with
UOHD. Most receivers realized that they could trust their vendors to conduct UOHD in a way that does not pose any major
risks to their establishments, or upset the local communities. The implementation in NYC followed a successful pilot test,
widely reported by the press (Journal of Commerce, 2009, 2010; Wall Street Journal, 2010), where a number of signature
companies participated. Some well-known companies that participated in the original pilot, and continued to receive
OHD, are Dunkin Donuts, Whole Foods, and Footlocker. Similar results have also been seen in other cities, such as Sao Paulo,
Brazil (Centro de Inovao em Sistemas Logsticos, 2015) and Bogota, Colombia (Universidad Nacional de Colombia Sede
Bogot, 2015) where successful pilots have also been conducted and steps have been taken to expand the pilots into full-
fledged OHD programs.
Furthermore, the behavioral research conducted by the authors (Holgun-Veras et al., in press-b) and the analyses of the
receivers that elected to use UOHD revealed that the availability of a Trusted Vendor (TV) significantly increases the recei-
vers willingness to accept UOHD. A TV was defined as one that the receiver would be comfortable allowing access to the
premises to deliver supplies during the off-hours without staff present to supervise the process. One such vendor that par-
ticipated in the original NYC pilot and expanded significantly its OHD operation is Sysco, one of the worlds largest distrib-
utor of food products, and the one with the largest number of OHD accounts to date among the participating vendors
(Holgun-Veras et al., 2011). The availability of a TV is important because one of the highest concerns for receivers as to
why they do not want to accept UOHDs is the (perceived and real) security risks of theft or damage involved in allowing
someone access to their property.
The importance of trust is hard to overestimate as in conditions where public safety is adequate and the economic con-
sequences of a negative outcome are low to moderate, the trustworthiness of the vendor becomes the critical factor. As in all
human relations, a receiver would learn to trust a vendor after an extended period of time of rewarding business or personal
relations. The Trusted Vendor Certification Program (TVCP) is a mechanism to expedite the process by which receivers
develop trust on a vendor. To this effect, the TVCP would certify vendors as trusted based on their qualifications and an
evaluation of their performance with UOHD. In this context, the TVCP would play a role similar to the one played by

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 3

customer reviews on Internet sale sites. The expectation is that the program will help diminish receivers concerns about
UOHD, ultimately increasing acceptance of UOHD.
Although there are solid reasons to believe that the TVCP will help foster adoption of UOHD, it is not clear what would be
the overall impact of the program, and what segments of the receiver population could be influenced by the TVCP. Solid esti-
mates of the expected impacts of the TVCP are crucial to decide whether or not the potential benefits justify the effort and
expense associated with its creation. Identifying the segments of receivers most inclined to participate is very important
because it enables outreach efforts to be focused on the most promising groups. The expectation is that once the TVCP takes
root in the most welcoming industry segments, it could spread its influence to others by helping assuage the concerns busi-
nesses from other sectors may have about UOHD. To this effect, the paper explores the acceptability of a TVCP among recei-
vers using discrete choice models estimated with the stated preference survey conducted by the authors. The insight gained
from the paper helps shed light into how best to implement and design a TVCP.
The paper is organized as follows. Section 2 provides background information on UOHD, and the concepts of a TV and the
TVCP. Section 3 conducts and report descriptive analyses of the sample. Section 4 introduces the methodology used in the
modeling process. Section 5 discusses the modeling results. Section 6 identifies policy implications, while Section 7 presents
the chief conclusions of the research conducted.

2. Background on trusted vendors and their influence on UOHD participation

The first indication of the important role played by the availability of a Trusted Vendor was found during the behavioral
research conducted during the design of the UOHD program in NYC. As part of that effort, the authors collected stated pref-
erence data to identify the most effective incentives for inducing receivers to accept UOHD. The survey posed questions to
receivers regarding their willingness to do UOHD in response to different policy scenarios. The nonmonetary factors consid-
ered in the scenarios were: providing (or not) business support services in exchange for participation in UOHD, providing
public recognition (or not) to the receivers that successfully implemented UOHD. Another question asked receivers if they
had a TV. The models estimated by the authors indicated that the availability of a TV was one of the most important factors in
the decision to accept UOHD. The estimates in (Holgun-Veras et al., in press-b) suggest that the establishments with a TV are
significantly more inclined to accept UOHD than other establishment that do not have a TV (39.17% vs. 27.30% for those that
do not have a TV). The importance of the TV is reflected in its subjective valuation, i.e., the amount of money that could pro-
duce the same effect as having a TV. The results showed that the availability of the TV had the highest subjective values rang-
ing from $1,741 to $36,538 (Holgun-Veras et al., in press-b), depending on the industry sector.
To further gain insight about what attributes define a TV, the authors conducted in-depth-interviews with fifteen of the
participants that responded to the trusted vendor question. The interviewees reported that some of the characteristics that
define a TV are: a reputable company name, a large company size, a long-term contract, a consistent delivery person that
they have built a relationship with over time, and conducting safe and on-time deliveries. The respondents were asked if
they would trust a vendor if it was certified by the TVCP. To this question nine of the fifteen (60%) responded positively with
yes or possibly. Quite significantly, two of the respondents that had indicated (in the survey) that they were not willing
to accept UOHD changed their opinions and indicated that they would accept UOHD from certified trusted vendors.
The idea is that a TVCP certifies vendors on the basis of: (1) the strength and coverage of the programs they have in place
to ensure safe and community friendly UOHD (e.g., driver training programs, use of low noise delivery technologies and prac-
tices, insurance coverage); and (2) customer reviews that summarize the receivers experiences with the vendors. Instead of
receivers solely relying on their individual experience with a vendor (which could take significant time to develop into a rela-
tionship of trust), receivers would benefit from the assessment of the vendor made by the TVCP and the collective experience
of multiple receivers to decide whether or not to accept UOHD from a vendor. A reputable TVCP is expected to spread the
word about the performance of certified vendors making it easier for receivers to hire these vendors and accept UOHD. This
will expedite the process by which a receiver concludes that a vendor is indeed a TV, effectively increasing the number of TVs
in the market. Since about 49% of receivers do not need further consent from anyone else to participate in UOHD (Holgun-
Veras et al., in press-b), a reputable TVCP that helps convince a portion of these receivers could lead to a significant increase
in UOHD and a concomitant drop of freight traffic in the day hours. Moreover, the evidence from the interviews indicate that
a reputable TVCP will make it easier for vendors, especially small ones, to be recognized as trustworthy by receivers.
Examining the role of trust in business transactions helps understand the potential of the TVCP. Dawes (2003) defines
professional trust as . . .the degree to which the people and organizations charged with developing and delivering a service
believe they can rely on the motives and predict the performance of the other participant(s). This definition is relevant to
the TVCP as its aim is to expedite the formation of trust to foster the continued implementation of UOHD. Trust is typically
built from the interaction between entities over a period of time. When there is a high level of trust the relationship is more
organic and informal, though well-understood, methods of interaction are formed. In contrast, if there is a low level of trust a
more formal framework is necessary to guide the relationship to similar results as if there was a high level of trust (Dawes,
2003). This dynamic of low trust is what the TVCP aims to address in order for UOHD to occur as there is a high security risk
for receivers to allow access to their property without the presence of the receiver. The TVCP aims to act as a formal frame-
work that will substitute for the time it typically takes to build a trusted relationship.

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
4 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

Halliburton and Poenaru (2010) looked into the consumers opinions and behaviors relating to trust in banking, insurance,
and mobile service providers. Their research established a number of concepts of relevance to the study of the relationship
between shipper/carrier and receiver in connection with the TVCP. They found that trust is of utmost importance to achieve
successful business relationships, especially those that have a high level of risk, uncertainty and vulnerability. UOHD could
generate some risk to receivers as they expose their properties to theft and damage; based on the study, trust works to
reduce this actual and perceived risk to the receiver. For the vendors providing the service, an additional benefit for having
trust is gaining customer loyalty.
The first key concepts discussed by Halliburton and Poenaru (2010) are that trust is built over time, and that trust min-
imizes risk. The next and most important concept is that customers assess the level of trust based on emotional trust and
rational trust. Rational trust is based on how well the business delivers its goods or services; while emotional trust is based
on the treatment of the customer beyond the profit motive. The dimensions of rational trust include: competence, knowl-
edge, dependability, ability, integrity, reliability, predictability, and credibility. The dimensions of emotional trust include:
empathy; feelings of security and perceived strength; personal experience and beliefs; benevolence/goodwill; and altruism.
The final concept is that trust is developed via a combination of multiple channels. In other words, the overall trust of the
business is impacted by customers interaction with front-line employees such as the sales people, marketing communica-
tions, and self-service technologies such as online accounts.
A study from the Economist Intelligence Unit, which looked on the role of trust in business collaboration, recommends
that specified trusted people be assigned to lead new collaborations. A consistent contact person that has already established
trust will continue to build and maintain a trusted relationship. This is important as an individual in a firm that is unable to
inspire trust in the partners could negatively influence the relationship. A high level trust breach such as illegal acts duplic-
ity, unauthorized sharing of information leads to immediate breach of the current project and stalling any further collabo-
rations. The study also finds that face-to-face communication is vital in building trust compared to phone or any virtual
communication (The Economist Intelligence Unit, 2008).
These considerations are central to the TVCP as it may include a certification component and a platform to provide and
see customer reviews. The combination of these components will offer various benefits. The certification side allows the
businesses to be endorsed and verified by an expert entity while the addition of the review platform allows customers to
provide reviews. Luca (2011) took a look at the effect of online customer reviews on the demand for restaurants using Yelp.-
com as a case study. Yelp.com provides a platform for customers to review businesses using a five star rating system and a
space where the customer can write about their experience. Each individual star rating from a reviewer is used to create the
average rating for the business which put by Luca, act as a simplifying heuristic to represent the quality of the restaurant as
individual reviews maybe too many to go through in detail to arrive at an assessment. The study found that an increase in
ratings of one star resulted in an increase in revenue of about 5.4%. This effect is driven by independent restaurants as the
opposite effect is occurring for chain restaurants which have seen a shift in revenue from chain restaurants to independent
ones. This is an example of the potential benefit of including the customer review component to the TVCP.
Online customer reviews are very important, as highlighted by a study done in 2015 looked at the impact of online
reviews on sales. The main points observed from the 2015 Local Consumer Review Survey were that 92% of those surveyed
read customer reviews for local business, an increase in 4% from the previous year (2014) and of 21% increase since 2010. The
92% that indicated that they read online reviews is split into 33% that read online reviews regularly and 59% that read them
occasionally. Eighty Percent (80%) of consumers trust online reviews as much as personal recommendations by varying
degrees, they trust online reviews: always (8%); if the reviews are perceived as authentic (31%); depending on business type
(22%); and, if multiple reviews are available (19%). This is an increase from 2010 of 13% and an increase of 1% from 2014;
which shows that over the years on-line reviews have grown to be trusted (Anderson, 2015).
This brief discussion clearly illustrates the importance of trust in business relations and the potential role that outside
review programs could play in fostering trust on a business venture. These insights suggest that at credible TVCP could foster
acceptance of UOHD.

3. Descriptive analyses of the sample data

To understand receivers attitudes toward the TVCP, a stated preference question was included in a survey to assess the
willingness of the respondent to accept UOHD if a vendor is certified by a TVCP. The sample has 450 observations; 62% (280)
are from the NYC metropolitan area (primarily New York City and Northern New Jersey) and 38% from the New York Capital
Region, NYCR, (Albany, NY and surrounding suburban areas).
Fig. 1 shows the approximate metropolitan areas covered by NYC metropolitan area and NYCR regions. NYC metropolitan
area is approximately 9000 sq. mi and has a population of nearly 20 million with nearly nine million employees. NYCR region
is nearly 3000 sq. mi and has a population of 0.9 million with about 0.42 million employees. As shown, there is a huge dif-
ference in the population density, per square mile between NYC (2184 people/sq. mi.) and NYCR (266 people/sq. mi).
The survey collected data about freight and service vehicle trips in the two regions displayed in Fig. 1. It consisted of three
blocks of questions that gather data pertaining to deliveries (cargo received) and shipments (cargo sent out), service trips,
and current operations and flexibility. Block A contained questions on deliveries and shipment made to and from the busi-
nesses, including: number of deliveries received, number shipments sent out, typical size and weight of deliveries and

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 5

Fig. 1. Map of New York City (NYC) and New York Capital Region (NYCR).

shipments, vehicle type used for both deliveries and shipments, types of commodities received and shipped, and who
transports the deliveries and shipments. Block B was the service trip section which inquired about the number of service
trips received, type of vehicle used in the service trip, most common types of planned and emergency service trips, percent-
age of planned and emergency service trips that occur during regular business hours and those that take place during off-
hours, and a willingness question to determine the interest in receiving planned service trips during the off-hours. The final
section, Block C contained questions pertaining to the businesses current operations and flexibility, such as: number of
employees at the establishment, both full-time and part-time; fleet owned; freight loading/unloading facilities present at
the business location; interest in consolidating shipments; and willingness to do UOHD with a vendor that has been certified
by a reputable organization as a trusted vendor which is the question at the core of this paper.
In order to increase efficiency of data collection, the respondents were not asked to respond to all sections of the survey.
Since the bulk of the freight trip generation is made by the freight intensive sectors (FIS), it was decided to collect more
observations from this group and reduce the number of observations from industries that are not freight intensive. However,
this reduction in the sample size were applied to Section A (320 observations) and Section B (290 observations) that focused

Table 1
Breakdown of number of observations in sample.

Sector Section-A Section-B Section-C

NYC NYCR NYC NYCR NYC NYCR
Freight Intensive Sectors (FIS) 179 118 88 58 179 118
Non-FIS 21 2 92 52 101 52
Total 200 120 180 110 280 170

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
6 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

on FTG. The stated preference question that was used for the modeling in this paper is from Section C that was posed to all
450 respondents in the sample. The breakdown of the sample is in Table 1.
The survey asked respondents to list the top three commodity types they routinely receive. The responses were coded
using the Standard Classification of Transported Goods (SCTG). For modeling purposes, the commodity types were coded
as binary variables and included as independent variables of the behavioral models tested. The mapping between industry
sectors and commodity types is analyzed later in the paper in connection with the modeling results.
The North American Industry Classification System (NAICS) was used to group the companies by industry categories at
the 2-digit level, as shown in Table 2. The results are broken down in two subgroups: (1) Freight-Intensive Sectors, or FIS,
which are the industry sectors where production and consumption of freight is the primary activity; and (2) Non-Freight-
Intensive Sectors, non-FIS, which are the sectors where the production and consumption of freight is of secondary impor-
tance. Of those surveyed, 64% of the NYC sample and 69% of the NYCR sample were from establishments that belong to a
FIS while the rest of the observations came from non-FIS establishments. The largest portion of the complete sample is man-
ufacturing (NAICS 31, 32, and 33), constituting 22% of the sample.
For the total sample, the delivery frequency ranged from less than 1500 deliveries/day with a mean of 18.85 deliveries/-
day, a standard deviation of 53.94 deliveries/day, and a median of 5 deliveries/day. The breakdown is shown in Table 3. It is
important to note that in Tables 3 and 4, the number of observations is less than the total sample (specified in Table 2) due to
nonresponses to the specified survey questions. Respondents that are listed as less than 1 (<1) delivery/day are those that
provided a weekly amount of deliveries that resulted in less than 1 when divided by 5.5 days. Approximately 93% receive
50 deliveries/day or less; with the majority of the responses being between 1 and 5 deliveries/day followed by the group
that receive 610 deliveries/day. Similar proportions were found in the separate NYC and NYCR samples. There are similar-
ities and differences between the responses provided by the FIS and non-FIS establishments. In both cases, the bulk of the
establishments receive less than 5 deliveries/day (64.4% of FIS, and 60.9% of non-FIS). However, the statistical distribution for
FIS establishments has a longer tail, with about 5% of the FIS of establishments receiving more than 75 deliveries/day. In con-
trast, no receivers in the non-FIS group received more than 50 deliveries/day.
The most important question for purposes of this paper is the one that asked respondents: If one of your vendors is
certified by a reputable organization to do deliveries in the night hours, how likely will you be to consider allowing that
vendor to make deliveries at night without any of your staff present? The range of answer options provided were: com-
pletely likely, very likely, neither likely nor unlikely, not likely, and completely unlikely. The responses are summarized in
Table 4. For discussion purposes, the responses corresponding to completely likely, are referred to in the paper as
committed; the combined response of completely likely and very likely are referred to as interested; while the com-
bined total of completely likely, very likely and neither likely, nor unlikely are assumed to represent the potential
market.

Table 2
Breakdown of sample by 2-digit NAICS.

NAICS Description NYC NYCR Total

Obs. % Obs. % Obs. %
Freight intensive sectors
23 Construction 20 11.17% 9 7.63% 29 9.76%
31 Manufacturing 10 5.59% 5 4.24% 15 5.05%
32 24 13.41% 16 13.56% 40 13.47%
33 27 15.08% 17 14.41% 44 14.81%
42 Wholesale trade 25 13.97% 30 25.42% 55 18.52%
44 Retail trade 21 11.73% 23 19.49% 44 14.81%
45 6 3.35% 5 4.24% 11 3.70%
48 Transportation and warehousing 16 8.94% 4 3.39% 20 6.73%
49 0 0.00% 1 0.85% 1 0.34%
72 Accommodation and food services 30 16.76% 8 6.78% 38 12.79%
Sub-total 179 100% 118 100% 297 100%

Non-freight intensive sectors

51 Information 13 12.87% 7 13.46% 20 13.07%
52 Finance and insurance 12 11.88% 2 3.85% 14 9.15%
53 Real estate and rental 8 7.92% 4 7.69% 12 7.84%
54 Professional, scientific, tech. services 8 7.92% 16 30.77% 24 15.69%
56 Administrative, support, waste 12 11.88% 11 21.15% 23 15.03%
61 Educational services 11 10.89% 4 7.69% 15 9.80%
62 Health care and social assistance 9 8.91% 6 11.54% 15 9.80%
71 Arts, entertainment, and recreation 21 20.79% 2 3.85% 23 15.03%
81 Other services (except public admin.) 7 6.93% 0 0.00% 7 4.58%
Sub-total 101 100% 52 100% 153 100%
Grand total 280 170 450

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 7

Table 3
Number of deliveries per day.

No. of deliveries/day NYC NYCR Total

Obs. % Obs. % Obs. %
Freight intensive sectors
<1 11 6.18% 17 14.53% 28 9.49%
15 82 46.07% 80 68.38% 162 54.92%
610 31 17.42% 9 7.69% 40 13.56%
1115 12 6.74% 5 4.27% 17 5.76%
1620 10 5.62% 1 0.85% 11 3.73%
2150 14 7.87% 3 2.56% 17 5.76%
5175 5 2.81% 0 0.00% 5 1.69%
76100 5 2.81% 2 1.71% 7 2.37%
>100 8 4.49% 0 0.00% 8 2.71%
Sub-total 178 100% 117 100% 295 100%

Non-freight intensive sectors

<1 1 4.76% 0 0.00% 1 4.35%
15 11 52.38% 2 100.00% 13 56.52%
610 3 14.29% 0 0.00% 3 13.04%
1115 2 9.52% 0 0.00% 2 8.70%
1620 3 14.29% 0 0.00% 3 13.04%
2150 1 4.76% 0 0.00% 1 4.35%
5175 0 0.00% 0 0.00% 0 0.00%
76100 0 0.00% 0 0.00% 0 0.00%
>100 0 0.00% 0 0.00% 0 0.00%
Sub-total 21 100% 2 100% 23 100%
Grand total 199 119 318

Table 4
Willingness to accept UOHD from a certified trusted vendor.

Willingness to accept NYC NYCR Total

Obs. % Obs. % Obs. %
Freight intensive sectors
Completely likely 9 5.14 4 3.45 13 4.47
Very likely 7 4.00 1 0.86 8 2.75
Neither likely, nor unlikely 8 4.57 3 2.59 11 3.78
Not likely 59 33.71 45 38.79 104 35.74
Completely unlikely 92 52.57 63 54.31 155 53.26
Sub-total 175 100 116 100 291 100

Non-freight intensive sectors

Completely likely 0 0.00 0 0.00 0 0.00
Very likely 5 5.21 1 1.96 6 4.08
Neither likely, nor unlikely 9 9.38 1 1.96 10 6.80
Not likely 39 40.63 23 45.10 62 42.18
Completely unlikely 43 44.79 26 50.98 69 46.94
Sub-total 96 100 51 100 147 100
Grand total 271 167 438

All sectors
Completely likely 9 3.32 4 2.40 13 2.97
Very likely 12 4.43 2 1.20 14 3.20
Neither likely, nor unlikely 17 6.27 4 2.40 21 4.79
Not likely 98 36.16 68 40.72 166 37.90
Completely unlikely 135 49.82 89 53.29 224 51.14
Total 271 100 167 100 438 100

The table shows that, overall, there is 2.97% of committed establishments, with another 6.17% interested in the program,
and a potential market of about 11% of the establishments. The remaining eighty-nine (89%) were not interested. In terms of
location, the responses are significantly different. NYC respondentsboth FIS and non-FISsuggest a larger potential market
(13.71% and 14.59%, respectively) than their counterparts in the Capital Region (6.90% and 3.92%). These results make sense
on account of the vastly different levels of congestion and their impacts on urban supply chains, and the general awareness

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
8 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

Table 5
Willingness to use a certified trusted vendor by industry sectors.

NAICS (Obs.) Description Committed Interested Potential Range

(completely likely) (completely/very likely) (completely/very/neutral)
72 (37) Accommodation and food services 13.51% 13.51% 16.22% 1416%
42 (53) Wholesale trade 5.66% 11.32% 13.21% 613%
4849 (20) Transportation and warehousing 5.00% 10.00% 10.00% 510%
4445 (55) Retail trade 3.64% 7.27% 12.73% 413%
3133 (97) Manufacturing 2.06% 3.09% 8.25% 28%
51 (20) Information 0.00% 5.00% 15.00% 015%
61 (14) Educational services 0.00% 7.14% 14.29% 014%
71 (22) Arts, entertainment, and recreation 0.00% 4.55% 13.64% 014%
54 (24) Professional, scientific, tech. services 0.00% 0.00% 12.50% 013%
52 (11) Finance and insurance 0.00% 9.09% 9.09% 09%
56 (22) Administrative, support, waste 0.00% 4.55% 9.09% 09%
53 (12) Real estate and rental 0.00% 0.00% 8.33% 08%
23 (29) Construction 0.00% 3.45% 6.90% 07%
62 (15) Health care and social assistance 0.00% 6.67% 6.67% 07%
81 (7) Other services (except public admin.) 0.00% 0.00% 0.00% 0.00%

of deliveries in the off-hours in NYC, among other factors. The results also show a difference between the responses from FIS
and non-FIS responders. While in the former case, about 4.5% of respondents indicated they were completely likely to accept
UOHD, in the latter none of the respondents expressed that level of interest. At the same time, the non-FIS group indicated a
significantly smaller number of responses in the completely unlikely group (46.94% vs. 53.26% in FIS responses) and a lar-
ger percentage of responses in neutral and not likely choices (6.80% and 42.18% vs. 3.78% and 35.74% in non-FIS). Essen-
tially, a larger portion of the non-FIS responses are clustered toward the middle of the ordinal scale (very likely, neither
likely nor unlikely, not likely), as the extreme positions (completely likely and completely unlikely are less significant
where 53.06% of non-FIS responses are found, vs. 42.37% of FIS.
The breakdown of the responses by industry sectors is shown in Table 5 for: (1) Committed (completely likely); (2)
Interested (completely likely plus very likely); and (3) Potential (completely likely plus very likely plus neither
likely, nor likely). Presenting these different estimates provide a sense about the varying degrees of certainty in the recei-
vers responses. To facilitate interpretation the authors decided to bold-italics the top five industry sectors in each set of
results. In cases where the difference between two sectors was less than 0.5% the authors assume a tie, and included both
industry sectors in the list. This is the reason why the last column has seven sectors in the top-five.
The results are very interesting. To start, the group of committed industry sectors is completely dominated by FIS as not
a single non-FIS makes the top-five list. Once the list is broadened to form the interested group, the non-FIS start to play an
important role. Broadening the list to include the neutral establishment, thus creating the group of potential adopters,
leads to a situation where about half of the top spots are occupied by non-FIS. These results have a number of significant
implications. First, they show that a sizable number of establishments could be induced to accept UOHD at the relatively
small investment of creating the TVCP. Second, the estimates clearly show that there is a potential market for UOHD in
the non-FIS group. This is important because non-FIS establishments do create a significant amount of deliveries, estimated
to be in excess of 250,000 deliveries/day in the NYC metropolitan area. The potential implications of these results are further
discussed later in the paper.

4. Methodology

To build on the insight gained from the descriptive analyses and get a better understanding of the data, discrete choice
models were estimated using the stated preference data. The benefits of using behavioral models include their ability to:
capture the influence of the agents attributes (e.g., industry sector, number of employees) and policy variables, e.g., tolls,
on the decision. In addition, these models provide a platform to analyze the agents responses to various policy alternatives
and to assess the respective market shares. The resulting estimates offer more detailed insight and offers valuable input in
policy making. Due to the ordinal nature of the dependent variable being estimated, an ordered choice model was used. The
dependent variable ranks the willingness to allow a TV that has been certified by a reputable program to perform UOHD. The
dependent variable is ranked from 1 to 5 in ascending order of willingness (i.e., 1 = completely unlikely, 2 = not likely,
3 = neither likely nor unlikely, 4 = very likely, and 5 = completely likely). The independent variables are attributes of the
receiving establishments.
The ordered choice models are based on the random utility theory where the utility (U) is derived as a function of inde-
pendent variables (X) as shown in the Eq. (1), where U is continuous and can take any real value. In ordinal utility models the
U can take particular values corresponding to each alternative. In this particular application the utility can take the five val-
ues representing the willingness to accept UOHD.
U bX e 1

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 9

where,
X: A vector of variables or attributes proper to the observation;
b: A vector of estimable parameters;
e: A random disturbance.

In case of ordinal models, the probability of an individual choosing an outcome i depends on the probability of U falling
between threshold parameters (li) as shows in Eq. (2). For n number of choices the model derives n  1 threshold
parameters.

P i Pli1 < U 6 li 2
The modeling process entailed systematic testing of alternative specifications of the systematic component of the utility
functions including interaction terms between the key independent variables. The resulting models were analyzed in terms
of statistical significance and conceptual validity; any model that did not meet these criteria was removed from considera-
tion. Once the final models were obtained, the individual elasticities of the key variables were calculated. Based on the esti-
mates for each individual, it is possible to compute the market elasticity to assess the overall effects of incentives (or
attributes) on the willingness to accept UOHD. The next section discusses the final models obtained during the estimation
process.

5. Model estimation results

Table 6 shows the model selected as final. After removing observations with missing data, the number of observations
included in the final model is 283 (170 observations from NYC and 113 from the NYCR). The measures of the goodness of
fit (McFadden Pseudo R2 and AIC) suggest that the controlled variables only slightly improve the prediction power, implying
great heterogeneity in business responses. In addition to the log-likelihood-based measures, measures based on model pre-
diction are also presented. Table 7 shows the cross tabulations of prediction (both outcomes and probabilities) versus the
actual outcomes, which yield additional measures of goodness of fit. The Count R2, which is the percentage of correct pre-
dictions, is 0.597, with low willingness level over-predicted. The adjusted Count R2 (Long and Freese, 2006) is 0.20.

Table 6
Discrete choice model estimating willingness to use a certified trusted vendor.

Independent variables Coefficient t-stat p-value 95% confidence

internal
Constant 0.025 (0.14) 0.892 0.335 0.386
Number of employees 0.004 (2.24) 0.025 0.008 0.001
Type of establishment
Headquarters 0.416 (1.47) 0.143 0.973 0.140
Commodity type
Food 0.760 (2.36) 0.018 0.128 1.393
Metal 0.577 (1.43) 0.152 1.366 0.213
Not a freight intensive sector
Interaction term for not being a FIS in the New York State Capital Region sample 1.415 (1.19) 0.233 0.911 3.741
Interaction terms: Number of deliveries (per day) and NAICS
Number of deliveries for entertainment (NAICS 71) 0.095 (1.35) 0.178 0.233 0.043
Interaction terms: Number of deliveries (per day) and commodity type
Number of deliveries of alcoholic beverages 0.127 (2.00) 0.046 0.003 0.251
Number of deliveries of plastics and rubber 0.045 (1.85) 0.065 0.003 0.092
Number of deliveries of wood and lumber 0.045 (1.86) 0.063 0.002 0.092
Number of deliveries of machinery 0.016 (1.16) 0.247 0.011 0.044
Number of deliveries of nonalcoholic beverages 0.014 (1.08) 0.281 0.011 0.039
Number of deliveries of computer and electronics 0.009 (2.95) 0.003 0.003 0.015
Number of deliveries of furniture 0.123 (1.85) 0.064 0.253 0.007
Number of deliveries of petrol and coal 0.132 (1.04) 0.299 0.382 0.117
Parameters
l(1) 2.222 (11.16) 0 1.832 2.613
l(2) 2.758 (11.33) 0 2.281 3.235
l(3) 3.357 (10.73) 0 2.744 3.970
Number of observations (n) 283
Log likelihood at convergence 278.08
McFadden Pseudo R-square 0.08
Akaike Information Criterion (AIC) 2.092

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
10 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

In interpreting the model it is important to keep in mind that retail and household goods were used as the default
base values for industry sector and commodity type, respectively. Thus, variables that contain either the industry sector
or a commodity type must be interpreted as marginal effects with respect to the corresponding base. Generic variables,
in contrast, apply to all establishments in the sample.
As shown, the willingness to accept UOHD from certified trusted vendorsreferred to herein simply as willingness to
accept or WTAis a function of employment, the type of commodity that is transported, number of deliveries, industry
sector (two digits NAICS), and whether or not the establishment is the headquarter of the firm. The model has a number
of noteworthy features. To start, the commodity type is found in ten out the thirteen independent variables, playing a
more significant role than the industry sector. This result is not entirely unexpected as the authors had long assumed that
the industry sector is only a rough proxy for the commodity type and other establishment-related attributes. Such proxy
cannot fully capture the effects associated with commodity type. For instance, retail establishments could receive a wide
range of supplies, from vegetables, all the way to jewelry. Obviously, the attitude toward UOHD and the TVCP form the
establishment owner will be heavily influenced by the nature of the supplies traded there. Nevertheless, the dominance of
the commodity typewhich replaced industry sector as the lead establishment descriptoris significant because it may
imply that mapping industry sectors to commodity types could increase the explanatory power of freight behavioral
models.
The number of employees was found to have a negative relationship with WTA across all establishments. The establish-
ment being the headquarter of the company also reduces willingness to participate, which makes sense because headquar-
ters typically have tight access controls to protect commercially sensitive information and any valuables stored there. The
results also indicate a significant difference between the establishments in non-freight-intensive-sectors located in the
NYCR, and the rest of the respondents. As shown in the table, these establishments exhibit a larger WTA for unassisted
off-hour deliveries performed by certified trusted vendors.
The model shows that the WTA of recipients of food is much higher than the one for the base industry sector (retail);
while the WTA for receivers of metal is significantly lower. Moreover, the commodity type has its most significant impact
interacting with the number of deliveries. As shown, there are two subgroups of establishments: those with WTA that
increases with the number of deliveries, and those that WTA that decreases with number of deliveries. The former group
includes receivers of alcoholic beverages, wood and lumber, plastics and rubber, machinery, non-alcoholic beverages, and
computer and electronics; while the latter group includes the receivers of furniture and petrol and coal. The important role
played by the commodity type does not imply that the industry sector plays no role. As shown in the model, the propensity
to participate diminishes with the number of deliveries to establishments in Entertainment (NAICS 71). It should be noted
that the test of parallel lines (Brant, 1990) indicated that the current model form, which assumes all parameters remain
consistent across all willingness levels, is too restrictive. The parallel lines test essentially compares the log-likelihood of
a model with level-specific parameters and that with invariant parameters across levels using a Chi-squared test.
Future research efforts should look into alternative model forms that allow parameters to vary across different willingness
levels.
The chief insight is that establishment that receive food, and particularly, those that receive large number of deliveries of
alcoholic beverages, wood and lumber, plastics and rubber, machinery, non-alcoholic beverages, and computer and electron-
ics represent the most welcoming environment to implement the TVCP. The policy implications of these findings are dis-
cussed next.

Table 7
Comparison between prediction and actual outcomes.

Willingness actual/predicted Completely likely Very likely Neither Not likely Completely unlikely Total
Cross tabulation of predictions and actual outcomes
Completely likely 1 0 0 5 6 12
Very likely 1 0 0 4 3 8
Neither 0 0 0 3 8 11
Not likely 0 0 0 29 73 102
Completely unlikely 0 0 0 16 134 150
Total 2 0 0 57 224 283

Cross tabulation of outcomes and predicted probabilities

Completely likely 1 0 1 4 6 12
Very likely 1 1 1 3 3 8
Neither 0 0 0 4 5 11
Not likely 5 3 5 40 49 102
Completely unlikely 4 3 5 49 89 150
Total 12 8 11 102 151 283

Note: Rows for actual outcomes and columns for predictions.

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 11

6. Policy implications

A chief policymaking principle is to ensure the maximum effectiveness of the effort. In the case considered in this paper,
the application of that principle means focusing on the industry sectors that offer the possibility of the largest payoff, in
terms of number of deliveries switched to UOHD. The latter is jointly determined by both the willingness of the industry
sector to switch to UOHD, and the Freight Traffic Generation (FTG)the summation of the freight trip production and attrac-
tionsof these establishments. The behavioral analyses of the previous section ought to be complemented with FTG esti-
mates to identify the best target industry sectors. Table 8 shows the FTG for the NYC metropolitan area, which was
estimated using a customized version of the Freight Trip Generation Estimator developed by the authors (Holgun Veras
et al., 2015), which uses the models from Holgun-Veras et al. (2012). The software uses publicly available data, i.e., employ-
ment distributions by postal codes (ZIP code), to compute FTG at the ZIP code level, by industry sector, and establishment
size (number of employees). Only NYC was chosen for the application because the authors feel that while the FTG models
could be applied to the NYCR sample as well, it is of more interest to focus the discussion on NYC as it is the more complex
case, because of its size and intricacies.
The table shows a number of significant results. To start, the retail, wholesale, and accommodation and food sectors gen-
erate in excess of 55% of the total FTG in the area. Moreover, the smallest establishments (less than five employees) generate
half the total FTG. Establishments in the retail, wholesale, and accommodation and food sectors with less than five employ-
ees generate almost a quarter (24%) of the total FTG. These estimates fly in the face of widely held views of large establish-
ments being the largest source of FTG. The reality is that, although large establishments do generate a large FTG per capita,
the more numerous small establishments generate as much FTG as all the other establishments combined.
To assess the range of impacts of the TVCP, the estimates of participation from Table 5 are combined with the FTG esti-
mates from Table 8 to create Table 9. Three different sets of estimates are provided, that depend on the assumption made
about participation in UOHD. The results indicate that if only the top-five sectors participate, between 3.94% and 8.89% of
deliveries could switch to UOHD. This would lead to a reduction of daytime freight traffic in the NYC metropolitan area
in the range of 54,000127,000 deliveries/day. If the entire potential market (11.19%) accepts UOHD, the reduction in the
number of deliveries during the day hours would increase to about 155,000 deliveries/day. These sizable numbers provide
strong justification to the study, design, and implementation of the TVCP.
The next step is to identify in more detail the establishment types offering the largest payoff. To this effect, the chief find-
ings of the model in Table 6that identified the most promising establishment typesand the FTG results of Table 8 were
put together to create Table 10. The table shows in the columns the commodity types, which have been sorted from left-to-
right in descending order of WTA. The last two columns represent the commodity types associated with opposition to the
idea of accepting UOHD. The rows show the various industry sectors, sorted out top-to-bottom in descending order of
FTG (at the level of two-digit NAICS). The upper left corner of the space below the commodity typesfor each of the NAICS
groups listedqualitatively represent the industry sectors that offer the largest payoff.
The table suggests that the most attractive targets for the TVCP are the industry sectors of: Food and Beverage Stores
(NAICS 445), Merchant Wholesalers/Nondurable Goods (NAICS 424), Food Services and Drinking Places (NAICS 722), Food

Table 8
Freight Traffic Generation (FTG) for the NYC metro area (000 deliveries/day).

NAICS Description Establishment employment

14 59 1019 2049 5099 100249 >250 Total %
44 Retail trade 177.9 63.6 46.6 49.5 26.7 27.5 22.7 414.44 29.9%
42 Wholesale trade 98.6 29.4 22.1 21.6 10.8 8.5 10.8 201.71 14.6%
72 Accommodation/food 59.7 26.2 25.5 32.7 15.4 7.1 4.7 171.21 12.4%
23 Construction 77.1 17.3 11.7 9.5 4.4 3.1 2.6 125.68 9.1%
31 Manufacturing 38.2 14.9 11.7 10.8 4.1 2.6 0.9 83.12 6.0%
54 Professional, scientific 51.8 8.7 5.1 3.2 1.1 0.7 0.3 71.00 5.1%
48 Transport/warehousing 34.2 8.2 6.1 5.7 3.2 3.4 4.0 64.81 4.7%
62 Health care/social 31.3 12.6 8.5 5.2 1.5 1.2 0.7 61.00 4.4%
81 Other services. . . 40.5 9.8 4.9 2.5 0.6 0.2 0.1 58.60 4.2%
53 Real estate/rental . . . 25.7 4.2 2.1 1.1 0.3 0.1 0.0 33.53 2.4%
52 Finance and insurance 16.4 6.0 4.1 1.8 0.7 0.5 0.3 29.78 2.2%
56 Administrative. . .waste 19.5 3.8 2.5 1.9 0.9 0.6 0.4 29.45 2.1%
51 Information 6.5 1.5 1.3 1.1 0.5 0.3 0.2 11.55 0.8%
71 Arts, entertainment. . . 7.5 1.2 1.0 1.0 0.4 0.2 0.1 11.42 0.8%
61 Educational services 4.3 1.3 1.1 1.3 0.6 0.3 0.2 9.05 0.7%
55 Management of. . . 1.1 0.3 0.4 0.5 0.3 0.3 0.2 3.12 0.2%
22 Utilities 0.9 0.3 0.3 0.3 0.3 0.2 0.1 2.53 0.2%
11 Agriculture, forestry. . . 1.4 0.2 0.1 0.0 0.0 1.68 0.1%
21 Mining, quarrying. . . 0.3 0.1 0.1 0.1 0.0 0.0 0.53 0.0%
Total 692.9 209.6 155.2 149.5 71.9 56.9 48.2 1384 100.0
% 50.1% 15.1% 11.2% 10.8% 5.2% 4.1% 3.5% 100.0%

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
12 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

Table 9
Participation in UOHD induced by TVCP by industry sector.

NAICS Description FTG% Committed Interested Potential Range

(completely likely) (completely/very likely) (completely/very/neutral)
72 Accommodation and food services 12.4% 13.51% 13.51% 16.22% 1416%
42 Wholesale trade 14.6% 5.66% 11.32% 13.21% 613%
48 Transportation and warehousing 4.7% 5.00% 10.00% 10.00% 510%
44 Retail trade 29.9% 3.64% 7.27% 12.73% 413%
31 Manufacturing 6.0% 2.06% 3.09% 8.25% 28%
23 Construction 9.1% 0.00% 3.45% 6.90% 07%
54 Professional/scientific/tech. services 5.1% 0.00% 0.00% 12.50% 013%
62 Health care and social assistance 4.4% 0.00% 6.67% 6.67% 07%
81 Other services (except public admin.) 4.2% 0.00% 0.00% 0.00% 0.00%
53 Real estate and rental 2.4% 0.00% 0.00% 8.33% 08%
52 Finance and insurance 2.2% 0.00% 9.09% 9.09% 09%
56 Administrative, support, waste 2.1% 0.00% 4.55% 9.09% 09%
51 Information 0.8% 0.00% 5.00% 15.00% 015%
71 Arts/entertainment/recreation 0.8% 0.00% 4.55% 13.64% 014%
61 Educational services 0.7% 0.00% 7.14% 14.29% 014%
55 Management of companies/enterprises 0.2%
22 Utilities 0.2%
11 Agriculture/forestry/fishing/hunting 0.1%
21 Mining/quarrying/oil and gas 0.0%
Total of top-five industry sectors 67.6% 3.94% 6.26% 8.89% 49%
Grand total (all industry sectors) 100.0 3.94% 7.18% 11.19% 411%

Services and Drinking Places (NAICS 722), and Accommodation (NAICS 721). It is advisable to add Food Manufacturing
(NAICS 311), and Beverage and Tobacco Product Manufacturing (NAICS 312) to the TVCP outreach efforts as doing so would
enable the public sector to focus on the entire food and beverage sector.
The finding that the Food and Beverage industryfrom food manufacturers to consumer oriented establishments in
accommodation and food servicescould be influenced by the TVCP is consistent with the experience with the UOHD imple-
mentation project that was conducted in NYC, where 4% of restaurants and food stores have already adopted UOHD. Focusing
outreach efforts in these sectors will enable policy makers to take advantage of the relations already developed with the pri-
vate sector. Moreover, the availability of willing private sector partners that could provide testimony of the real-life benefits
of UOHD will assist the efforts of designing and implementing the desired trusted vendor certification program, or TVCP.
It is important to highlight that these encouraging results do not imply that the TVCP could replace the use of incentives
as the chief mechanism to foster UOHD. For instance, a fully-fledged UOHD program is estimated to induce about 40% of the
receivers in the food and accommodation sectors to accept UOHD. This number is about three times larger than what could
be accomplished with the TVCP alone. However, the authors believe that the TVCP is extremely cost-effective.
In the authors opinion the TVCP could be a very powerful synergistic complement to an incentive-based UOHD program,
and proactive private sector involvement. In this view, the TVCPs role would be to help connect receivers with TVs, and pro-
vide honest information to the market about the vendors. The key function of the incentive program is to provide a nudge to
the receivers in the direction of accepting UOHD. The role of the private sector is crucial as they are the ones that have the
responsibility of making UOHD work, in terms of providing the quality of service needed by the receiver, and ensuring that
the quality of life of the surrounding communities is not negatively impacted. By helping spread the word about the real-life
experience with UOHD, the TVCP will help reduce the perception of risk among receivers. In doing so, it will help lower the
amount of the financial incentives needed to convince receivers to accept UOHD.

7. Conclusions

The research reported in this paper has studied the role and potential impacts that a Trusted Vendor Certification Program
(TVCP) could have in inducing receivers to accept unassisted off-hour deliveries (UOHD). The TVCP would provide informa-
tion to receivers about the track record of vendors, by means of a certification that attest that the vendor has attained the
status of a trusted vendor (TV). The TVCP will certify that the vendor has a good track record on conducting UOHD in a way
that does not put the establishment at risk, in accordance with applicable regulations, and without generating negative
impacts on the surrounding communities. A reputable TVCP would make it easier for receivers to find trusted vendors. In
doing so, the TVCP is increasing the level of transparency of the market with the likely effect of increasing participation
in UOHD.
To assess the potential impacts and the industry sectors most inclined to accept UOHD in response to the TVCP, the
authors conducted a survey with 450 observations collected from the New York City (NYC) metropolitan area, and the
New York Capital Region (NYCR). To quantify the expected market share and identify the number of deliveries that could
be impacted by the TVCP, the authors used the survey results, the insight gained from the ordinal choice models estimated

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 13

Table 10
Top three commodities received by industry sector.

NAICS NAICS description (top to Commodities (left to right, descending order of WTA)
bottom, descending order
Food Alcoholic Non- Plastics Wood Machinery Computer Petrol Furniture Metals
of FTG)
beverages alcoholic & & & & coal
beverages rubber lumber electronics
445 Food and beverage stores U U U U U
441 Motor vehicle and parts U U U U
dealers
442 Furniture, home U U U
furnishings stores
453 Miscellaneous store U U U U
retailers
444 Building material, garden U U
equip. . .
443 Electronics and appliance U U
stores
446 Health and personal care U
stores
452 General merchandise U
stores
424 Merchant wholesalers, U U U U
nondurable. . .
423 Merchant wholesalers, U U U U U
durable . . .
454 Nonstore retailers U
722 Food services and drinking U U U U U
places
721 Accommodation U U U
237 Heavy and civil U U
engineering const.
236 Construction of buildings U U U
238 Specialty trade U U U U U U
contractors
311 Food manufacturing U U U
312 Beverage/tobacco product U U
manuf.
325 Chemical manufacturing U U U U U
326 Plastics and rubber U U
products manuf.
327 Nonmetallic mineral U U
product manuf.
331 Primary metal manuf. U U U
332 Fabricated metal product U U U
manuf.
333 Machinery manufacturing U U U U U
334 Computer and electronic U U U
product. . .
335 Electrical equipment, U U U U
appliance. . .
339 Miscellaneous manuf. U U U U U U
321 Wood product U
manufacturing
337 Furniture and related U
product manuf.
313 Textile mills U U
323 Printing and related U U
support activities
324 Petroleum and coal U
products manuf.
336 Transportation equipment U U
manuf.
484 Truck transportation U U U U U U
493 Warehousing and storage U U U
488 Support activities for U U U U
transportation
485 Transit, ground passenger U
transp.

(continued on next page)

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
14 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx

Table 10 (continued)

NAICS NAICS description (top to Commodities (left to right, descending order of WTA)
bottom, descending order
Food Alcoholic Non- Plastics Wood Machinery Computer Petrol Furniture Metals
of FTG)
beverages alcoholic & & & & coal
beverages rubber lumber electronics
711 Performing arts, spectator U U U U U U
sports. . .
713 Amusement, gambling, U U U U U
recreation. . .
712 Museums, historical sites, U U
...

with the behavioral data collected in the survey, and freight trip generation estimates. The innovative used of this holistic
approach clearly indicated that the TVCP could have a significant effect on receivers acceptance of UOHD. The estimates
are that impacts produced by the TVCP ranges, in terms of number of deliveries switched to UOHD, range from 3.94% (if
all respondents that indicated they were completely likely to switch to UOHD do so), to 7.18% (if the completely likely
and very likely groups agree to accept UOHD), to 11.19% (if the respondents in the categories of completely likely, very
likely, and neither likely, nor unlikely decide to participate). The significance of these impacts is obvious if one takes into
account the TVCP could increase UOHD by simply increasing the level of transparency of the market.
The behavioral models estimated that willingness-to-accept (WTA) unassisted off-hour deliveries from certified trusted
vendors depends on a series of establishment attributes. Both establishment employment and the establishment being the
headquarters, have a negative relationship with WTA. Essentially, large establishments and headquarters are not good can-
didates for this program. The number of deliveries received plays a role that depend on the commodity type being delivered.
In the case of receivers of alcoholic beverages, wood and lumber, plastics and rubber, machinery, non-alcoholic beverages,
and computer and electronics, the WTA increases with the number of deliveries. In contrast, WTA decreases with the number
of deliveries of furniture, and petrol and coal. Receivers of food are inclined to participate, while receivers of metal are not.
It should be noted that, in spite of the encouraging research results discussed in the paper, it is important to pilot test the
TVCP concept. Doing so would provide additional evidence concerning the validity of the behavioral findings obtained from
the research, and will generate further insight into the real life acceptance of the TVCP. The insight gained is likely to play a
key role in enhancing the effectiveness of the TVCP.
To reach conclusions regarding which industry sectors offer the largest potential payoffin terms of deliveries switched
to the off-hoursthe authors conducted freight trip generation analyses and compared the resulting estimates to the sectors
identified by the behavioral models as the most inclined to participate. This analysis led to the conclusion that food and bev-
erage industry, and wholesalers of non-durable goods, offer the greatest potential payoff. Focusing on the food and beverage
sector is consistent with the experience of the UOHD implementation project that was conducted in NYC, which has
switched more than 400 establishments to the off-hours. Building on that track record, and the strong relations already
developed with the private sector in the food/beverage sectors seem to offer the highest probability of success.
In closing, it is important to note that the TVCP is best seen as a synergistic component of a tripod where the TVCP helps
connect receivers with trusted vendors, an incentive-based UOHD program provides a nudge to receivers to induce them to
try UOHD, the private sector vendors and their carriers focus their energies on conducting UOHD in a way that meets the
needs of the receivers and ensures that the local communities are not negatively impacted. Such holistic approach provides
the best chances of implementing an UOHD strategy that is beneficial to all.

Acknowledgements

The research reported in the paper was primarily funded by the New York State Energy Research and Development
Agencys project C-14-10 Off Hour Delivery Trusted Vendor Program and partially funded by the National Cooperative
Freight Research Cooperative Program projects NCFRP 25-01 Freight Trip Generation and Land Use, Strategic Highway
Research Program 2s SHRP2 C-20 Implementation Assistance Program for Freight Data Modelling and Innovation, and by
the VREF Center of Excellence for Sustainable Urban Freight Systems. Their support is both acknowledged and appreciated.

References

Anderson, M., 2015. 87% of Potential Customers Wont Consider Businesses with Low Ratings. Retrieved 5/19/2016, 2016, from <http://
searchengineland.com/87-percent-customers-wont-consider-low-ratings-228607>.
Brant, R., 1990. Assessing proportionality in the proportional odds model for ordinal logistic regression. Biometrics 46 (4), 11711178. http://dx.doi.org/
10.2307/2532457.
Brom, M.A., Holgun-Veras, J., Hodge, S., 2011. Off-hour deliveries in Manhattan, New York City: experiences of pilot test participants. Transp. Res. Rec. 2238,
7785. http://dx.doi.org/10.3141/2238-10.
Centro de Inovao em Sistemas Logsticos, 2015. Avaliao do Projeto-Piloto de Entregas Noturnas no Municpio de So Paulo. Serie Cadernos Tecnicos, 18,
from <http://www.antp.org.br/website/biblioteca/>.

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011
 J. Holgun-Veras et al. / Transportation Research Part A xxx (2016) xxxxxx 15

Dawes, S., 2003. The Role of Trust in New Models of Collaboration. New Models of Collaboration A Guide for Managers, from <https://www.ctg.albany.edu/
publications/online/new_models/essays/trust.pdf>.
Halliburton, C., Poenaru, A., 2010. The Role of Trust in Consumer Relationships. From <http://news.pb.com/white-papers/the-role-of-trust-in-consumer-
relationshipsescp.download>.
Holgun-Veras, J., 2011. Urban delivery industry response to cordon pricing, time-distance pricing, and carrier-receiver policies in competitive markets.
Transport. Res. Part A 45, 802824.
Holgun-Veras, J., Encarnacin, T., Gonzlez-Caldern, C.A., Winebrake, J., Kyle, S., Herazo-Padilla, N., Kalahasthi, L., Adarme, W., Cantillo, V., Yoshizaki, H.,
Garrido, R.A., in press-a. Direct impacts of off-hour deliveries on urban freight emissions. Transport. Res. Part D: Transp. Environ. (in press-a)
Holgun-Veras, J., Wang, C., Snchez-Daz, I., Campbell, S., Hodge, S.D., Jaller, M., Wojtowicz, J., inpress-b. Fostering unassisted off-hour deliveries: the role of
incentives. Tranport. Res. Part A: Pol. Pract. (in press-b)
Holgun-Veras, J., Jaller, M., Sanchez-Daz, I., Wojtowicz, J.M., Campbell, S., Levinson, H.S., Lawson, C.T., Powers, E., Tavasszy, L., 2012. Freight Trip Generation
and Land Use. National Cooperative Highway Research Program/National Cooperative Freight Research Program Retrieved NCHRP Report 739/NCFRP
Report 19, from <http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_rpt_739.pdf>.
Holgun-Veras, J., Ozbay, K., Kornhauser, A.L., Brom, M., Iyer, S., Yushimito, W., Ukkusuri, S., Allen, B., Silas, M., 2011. Overall impacts of off-hour delivery
programs in the New York City metropolitan area. Transp. Res. Rec. 2238, 6876.
Holgun-Veras, J., Silas, M.A., Polimeni, J., Cruz, B., 2007. An Investigation on the effectiveness of joint receivercarrier policies to increase truck traffic in the
off-peak hours: Part I: the behaviors of receivers. Netw. Spatial Econ. 7 (3), 277295. http://dx.doi.org/10.1007/s11067-006-9002-7.
Holgun-Veras, J., Silas, M.A., Polimeni, J., Cruz, B., 2008. An Investigation on the effectiveness of joint receiver-carrier policies to increase truck traffic in the
off-peak hours: Part II: the behaviors of carriers. Netw. Spatial Econ. 8 (4), 327354. http://dx.doi.org/10.1007/s11067-006-9011-6.
Holgun Veras, J., Aros-Vera, F., Wojtowicz, J., 2015. Freight Trip Generation Estimator: Improving Freight System Performance in Metropolitan Areas.
Journal of Commerce (2009, September 7, 2009). New York delivers at night. J. Commer. 10(35) pp. 4A.
Journal of Commerce (2010, July 2, 2010). New York to expand off-peak truck program. J. Commer. (July 3, 2010).
Long, J.S., Freese, J., 2006. Regression Models for Categorical Dependent Variables Using Stata. Stata Press, College Station, TX.
Luca, M., 2011. Reviews, Reputation, and Revenue: The Case of Yelp.com, Harvard Business School Negotiations, Organizations & Markets Unit Working
Paper No. 12-016.
Samuelson, P.A., 1952. Spatial price equilibrium and linear programming. Am. Econ. Rev. 42 (3), 283303.
Taniguchi, E., Noritake, M., Yamada, T., Izumitani, T., 1999. Optimal size and location planning of public logistics terminals. Transport. Res. Part E: Logist.
Transport. Rev. 35 (3), 207222.
The Economist Intelligence Unit, 2008. The Role of Trust in Business Collaboration An Economist Intelligence Unit briefing paper sponsored by Cisco
Systems from <http://graphics.eiu.com/upload/cisco_trust.pdf>.
Universidad Nacional de Colombia Sede Bogot, 2015. Desarrollo de una consultora para disear e implementar un piloto de cargue y descargue nocturno
en empresas de dos unidades de planeamiento zonal (UPZ) de Bogot. Resumen Ejecutivo.
Wall Street Journal, 2010. Congestion Fight Runs Into Night. Retrieved July 2, 2010, from <http://online.wsj.com/article/
SB10001424052748704334604575339292960610492.html>.

Please cite this article in press as: Holgun-Veras, J., et al. Role and potential of a trusted vendor certification program to foster adoption of
unassisted off-hour deliveries. Transport. Res. Part A (2016), http://dx.doi.org/10.1016/j.tra.2016.09.011