“Information and communication technologies, road freight transport, and environmental sustainability”

Despite progress in reducing air pollutants in several countries, freight transport con- tinues to have undesirable effects on environmental quality, human health, and the economy. Road freight transport, in particular, is associated with various negative ex- ternalities, including environmental and health damages, and the overexploitation of non-renewable natural resources. This paper investigates how ICTs interact with road freight transport to affect environmental quality regarding reducing CO 2 emissions. The empirical strategy is focused on the yearly dataset from 2002 to 2014 in 43 coun- tries. Using the two-step GMM techniques, the findings suggest that ICTs can decrease road freight transport’s negative impacts on environmental sustainability. Besides, the interactions of mobile phone and fixed telephone technologies with road freight transport are more efficient in reducing pollution than using internet networks. This paper underlines the importance of using ICTs to dampen road freight transport’s negative effects on environmental sustainability.


INTRODUCTION
Transportation is a major contributor to global CO 2 emissions, one of the main factors in global warming. If the average global temperature exceeds 2° C's safety level, then catastrophic impacts on environmental sustainability could ensue (IPCC, 2014). Transport contributed 23% of global carbon emissions from energy consumption, whereas road transport contributed 20% in 2014 (International Energy Agency, 2016). For example, road transport represents almost 92% of total CO 2 in UK, and 6% of total CO 2 emissions (Piecyk & McKinnon, 2010). Since the ratification of the Kyoto Protocol on Climate Change, CO 2 emissions have increased, especially in some regions such as China and India, while decreasing in certain areas, such as Europe (Olivier & Muntean, 2014). Despite significant decreases in many other sectors of the economy, CO 2 emissions have continued to increase in the transportation sector, and reducing environmental damage caused by transport seems to be more costly than in other sectors because it heavily depends on fossil fuels and traditional supporting infrastructure (i.e., diesel and petrol used in combustion engines). Consequently, all involved actors and users of road transport services have attempted to integrate some innovative solutions and clean technologies to reduce energy consumption by transportation and environmental pollution.
For reducing CO 2 emissions and associated costs of their road freight transport networks, many companies in recent years have adopted rapid advances in ICTs and their useful applications for road transport. These innovative applications can include on-vehicle systems (i.e., telematics and radio frequency identification) and in-house systems (e.g., computerized vehicle scheduling system). In the same context, the adoption of useful internet applications (e.g., electronic marketplaces, e-tickets, software systems, smart transport systems, and electronic reservation) is becoming very popular nowadays. These can also help companies identify the best combinations of routes that consume less energy and stimulate good driving practices to improve environmental quality.
This study aims to examine how ICTs can diminish the potential negative effect of road freight transport on the environment and identify the most efficient ICT technologies (internet, telephone, and mobile phones) that can be used to reduce CO 2 emissions. The main contributions are twofold: theoretically speaking, this study aims to develop the knowledge about the role played by the interaction of ICTs with road freight transport in reducing pollution, and from an empirical point of view, this study proposes some practical options to accomplish environmental benefits using some specific ICTs devices.

LITERATURE REVIEW
Although there is an important role to be played by ICTs in increasing cost efficiency and reducing pollution, very few studies have attempted to explore potentials in CO 2 reduction by adopting ICT for road freight transport ( Perego, Perotti, and Mangiaracina (2011) in their study identifying relationships between using ICTs for logistics and freight transport. Considering more than 40 papers, they found only the paper of Button, Doyle, and Stough (2001) that highlighted the impact of using ICTs in the context of freight transport to reduce environmental degradation. To our knowledge, few articles have been using an empirical methodology to demonstrate the impacts of ICTs in reducing environmental damages of road freight transport. Regarding the first relation, Añón Higón, Gholami, and Shirazi (2017) studied the relationship between ICT and environmental pollution. They mentioned that the ICT turning level is located above the mean value for developing countries, but the opposite scheme is true for developed countries. These results mentioned that several developed countries have already reached a level of ICT development whereby pollution is decreased. Asongu, Le Roux, and Biekpe (2018) examined how ICTs can influence Africa's environmental sustainability using a set of 42 countries between 2000 and 2012. They showed a positive relationship between ICTs and CO 2 emissions due to energy consumption. Asongu (2018) investigated how ICTs interact with globalization to affect CO 2 emissions using panel data of 44 countries from 2000 to 2012. He showed that ICTs could decrease the negative effects of globalization by reducing environmental damage.
Danish et al. (2018) also investigated how ICTs, per capita GDP growth, and financial expansion influence environmental quality. They considered two different associations: between ICTs and GDP per capita, and between ICTs and financial development. They employed an empirical strategy using a set of econometric techniques over the period 1990-2015. They concluded that ICTs, financial development, and per capita GDP growth positively affect CO 2 emissions. When ICTs are associated with financial development, this may increase environmental damage. The only positive effect on environmental sustainability is the interaction between ICT and per capita GDP growth. Moreover, Park et al. (2018) investigated the potential impacts of new technologies on the pollution reduction for 23 EU countries between 2001 and 2014, confirming that ICTs may positively in-crease carbon emissions in the long run. Similarly, the usage of electricity consumption has the same positive effect. However, per capita GDP growth, trade, and financial expansion negatively affect carbon emissions. In terms of policy implications, they recommended that the EU countries encourage ecological ICT devices and adopt the new inventions on the internet to decrease environmental degradation.
To understand the empirical relationship between transportation and carbon emissions, Saidi and Hammami (2017) considered 75 countries between 2000 and 2014 using three sets of countries categorized according to income level. They employed dynamic models in a simultaneous system of equations estimated through GMM procedure, consequently detecting a unique directional causality going from freight transport to CO 2 emissions for the high-income panel. Moreover, they found a negative effect of road freight transport on low-and middle-income countries' environmental degradation. Li and Yu (2017) investigated how smartphones could interact with road freight transport to reduce environmental damage. For this, they used some case studies to evaluate their applications in China. By conducting some interviews and using other data collection techniques, they found that freight applications can help identify the adequate trucks, achieve the delivery process, and improve traffic management. Therefore, this can improve efficiency and decrease CO 2 emissions from fuel consumption.
Other studies underscore the imperative necessity to manage logistics operations positively influencing environmental quality, especially given that 6% of net carbon emissions are from road freight transport (McKinnon, 2010). The rapid growth in the amount of energy consumption for road freight transport surpasses the total quantity of energy used by cars and buses (Wang et al., 2015). Therefore, some technical solutions are proposed by several studies to reduce pollution coming from road freight networks. In the same context, the European Commission investigated how ICT can reduce energy consumption 1 It is worth noting that there is no available data for the variable CO2 emissions from liquid fuel consumption from 2015 till now. Consequently, the selected data is limited to the period between 2002 and 2014. However, this limitation does not affect the quality of estimations and results.  Table 1 indicates the full variable definitions and related data sources. Table 2 shows descriptive data. Table 3 identifies the correlations that can exist between all variables used in the empirical investigation. According to Brambor, Clark, and Golder (2006), the existence of multicollinearity issues is considered irrelevant in interactive regressions.

Empirical strategy
To investigate how ICTs interact with road freight transport to influence environmental quality, a two-step GMM empirical strategy can be adopted. This choice is due to five main reasons as cited by Asongu (2018)    Note: The estimated p-values are in brackets. *** Significant at 1%. ** Significant at 5%. * Significant at 10%. 9353); (5) since the GMM approach is focused on panel data, disparities between countries are considered in regressions.
The Arellano and Bover (1995) extension developed by Roodman (2009) 3 is used since it can easily control the increase of instruments, and thus consider the dependence that can exist between sections 4 (Boateng, Asongu, Akamavi, & Tchamyou, 2016). The standard system GMM is represented by the following equations in level and first difference (respectively), in which the error term takes a two-way error component form:   Table 4 shows the main findings. It considers two different specifications: the first one is without control variables, while the second contains four control variables. In each specification, three subsets of specifications related to road freight transport are presented. Besides, each sub-empirical specification contains three different measures of ICTs (i.e., mobile phone, internet, and telephone). Two tests are exploited to ensure that the GMM procedure is appropriate: the "test of overidentifying restriction" and the "test of endogeneity/exogeneity". The Hansen J-test outcomes appear in Table 4, in which H0 allows us to confirm that the instruments are not correlated with the error term, and the excluded instruments are not considered in the estimation.

RESULTS AND DISCUSSION
This study also reports the Arellano and Bond (1991) test, namely AR(2), where the null hypothesis indicates that the differenced errors are auto-correlated since the regression errors are not dependent and equally distributed. Albeit the Sargan test is not robust, this corresponding test is not negatively affected by instruments, contrary to the Hansen test. To address this conflict, the Hansen test can be used for controlling the increase of instruments. 5 The results of the Hansen J-test and AR(2) are shown in Table 4. The first test approves the acceptance of H0, which means that the instruments are suitable, while the AR(2) test illustrates no autocorrelation evidence.
The findings broadly show the capability of ICTs in reducing the costs and constraints associated with freight transport using road transport networks, thereby decreasing CO 2 emissions and associated environmental degradation. This impor- 5 It is possible to avoid the proliferation of instruments when they do not exceed the number of groups in chosen sub-empirical specifications.
tant result is confirmed in most empirical studies (Baumgartner et al., 2008;Wang et al., 2015). More precisely, it appears that the interaction between ICTs (e.g., internet, mobile phone, and telephone penetration) and road freight transport has significant negative effects on CO 2 emissions, and therefore positive effects on environmental sustainability. Besides, relative to the specifications without control variables, the association of telephone penetration with road freight transport ( ) TEL FT ⋅ seems to be more efficient in reducing CO 2 emissions than the internet and mobile phone penetration. The magnitude of -0.226 means that a 10% improvement in TEL FT ⋅ may

CONCLUSION
This study's main goal is to investigate how ICTs interact with road freight transport to reduce carbon emissions in 43 countries over the period 2002-2014. The chosen period is conditioned by data availability, especially for the dependent variable (i.e., CO 2 emissions from liquid fuel consumption). The dataset is extracted from the World Development Indicators (World Bank). It is worth noting that this limitation does not affect the quality of estimates. The findings provide significant results concerning the impact of using ICTs in road freight transport on environmental sustainability. Although ICTs can play an important role in decreasing road transport's potential negative effect on environmental quality, few research contributions have been made. Using the two-step GMM techniques, the findings mainly demonstrate the ability of ICTs in decreasing potentially negative environmental impacts of road freight transport (i.e., reducing carbon emissions). The negative relationships between the use of ICTs (i.e., telephone, internet, and mobile phones) in road freight transport and CO 2 emissions attest to those technologies' capability in decreasing the costs and constraints related to the freight transport activity.
For both specifications, the impact of using ICTs in road freight transport on environmental quality was positive, regarding the reduction of CO 2 emissions. Indeed, a negative and significant impact was estimated for the interaction (TEL*F T ), where a 10% increase in TEL*F T may decrease CO 2 emissions by 2.26%. Moreover, for the interactions (MOB*FT) and (INT*FT), a 10% increase in MOB*FT and INT*FT may reduce carbon emissions by 0.85% and 0.74%, respectively. Thus, we conclude that the interaction (TEL*F T ) can decrease CO 2 emissions more efficiently than the other interactions, (MOB*FT) and (INT*FT). However, when considering the empirical specification with a control set of variables, it appears that the association (MOB*FT) becomes the most efficient, leading to an increase in environmental quality. A 10% increase in MOB*FT may reduce CO 2 emissions by 1.59%.
Finally, this paper is the first to explicitly propose an empirical analysis to identify the ability to use ICTs in damping the negative effects of road freight transport on environmental quality. This paper suggests applying ICTs in freight transport to control transport convoy, fuel use, pollution, and efficiency of the traffic system. Another benefit of using ICTs in the transportation sector is limiting the need for physical contacts and, thus, the urban costs (e.g., commuting costs, delays, accidents, etc.).