Factors Influencing Greenhouse Gas Reduction Measures in European Ports: Implications for Sustainable Investing

Abstract

European Union cargo and container ports are under pressure to reduce GHG emissions and achieve carbon neutrality by 2050, as mandated by the European Commission. The pace of progress varies among ports. This study examined the characteristics influencing GHG reduction measures in European cargo and container ports and their implications for sustainable investing. The methods used in this study, such as linear regression models to analyze predictive variables, can be applied in sustainable investing to assess which factors most strongly predict a company’s environmental, social, and governance performance. Using linear regression models to analyze data from the 33 busiest European ports, we identified five predictive variables: port size, cargo mix, surrounding population density, access to the sea, and the economic wealth of the host country. Our findings revealed that the port size significantly correlates with the adoption of measures to reduce scope 1, 2, and 3 emissions. This study underscores the importance of contextual and operational factors in evaluating sustainability efforts across sectors. The results contribute to drawing parallels with the field of sustainable investing within finance. This offers valuable insights for sustainable investing, emphasizing the importance of considering various contextual and operational factors when evaluating the sustainability efforts of entities in different sectors.

1. Introduction

Ports are facing stakeholder (Giuliano and Linder 2013; Sornn-Friese et al. 2021) and legal pressure to become greener (European Commission 2019). The top 10 most-polluting European ports emitted 44 million tons of CO₂ in 2018 (Transport and Environment 2022). This highlights the significant role of European ports in contributing to the GHG emissions within the maritime trade. As the global maritime trade is expected to keep increasing in the coming years without additional GHG reduction measures, GHG emissions will increase in maritime transport by 90–130% by 2050 compared to 2008 (IMO 2021).

The European Commission set the objective through the European Green Deal of becoming the first carbon-neutral continent by 2050, to mitigate the effect of climate change. This implies that all sectors will decrease their GHG emissions by at least 55% by 2030 compared to 1990 (European Commission 2021). The majority of European ports are already experiencing the effects of climate change from rising sea levels and extreme conditions such as erosion (ESPO 2021; Deloitte and ESPO 2021). Ports are a central node in international shipping for connecting maritime transport to the hinterland. Ports thus have a huge role to play in reducing the GHG emissions from international shipping (ESPO 2021; Deloitte and ESPO 2021; Gibbs et al. 2014). Not only they can reduce the direct GHG emissions of their activities, but they can also impact the GHG emissions of sea and inland transport. For instance, ports can implement solutions that guide cargo owners to reduce their emissions and thereby make the use of water transport and freight more attractive for the hinterland (Gonzalez-Aregall et al. 2021; Gonzalez-Aregall et al. 2018).

In recent years, European ports have become more concerned about climate change. It was added to the list of the top 10 environmental priorities of European ports in 2017 and became the first environmental priority in 2021 (ESPO 2022a, 2022b).

Maritime organizations such as the IAPH (International Association of Ports and Harbors), through the WPCI (World Ports Climate Initiative) launched in 2008, and the ESPO (European Sea Ports Organisation) promote green measures to be implanted by ports, such as monitoring GHG emissions, using on-shore power, electrifying port activities, and using renewable energy.

In the last decade, port authorities have widened their scope of activity by increasingly becoming hubs of energy, industry, and a blue and circular economy (ESPO 2021). European ports are also implementing measures to reduce their GHG emissions (ESPO 2022a, 2022b). In 2022, 63% of European ports were monitoring their carbon footprint (+15 percentage points compared to 2013), and 55% were providing OSP in at least one berth (a 71% increase compared to 2015) (ESPO 2022a, 2022b). However, not all European port are moving at the same pace (ESPO 2021). Port differs vastly in their characteristics (Puig et al. 2015; DNV-GL 2021) in term of size, governance, access to the sea, sector of activity, country of location, and population density of the surrounding area. Each characteristic might have an influence on the adoption by the port of green measures to reduce its GHG emissions.

Even if the literature about green port measures to reduce GHG emissions is rapidly growing (Lin et al. 2022), few research studies have focused on assessing the correlation between the port characteristics and the adoption of GHG reduction measures in the European Union socio-political context (Puig et al. 2015; DeSombre et al. 2023). According to Wang et al. (2023), understanding the specific carbon emission sources and technical measures for emission reduction at ports is crucial for tailoring effective strategies. Song (2024) emphasizes the importance of comprehensive decarbonization measures and roadmaps tailored to the unique operational dynamics of seaports, highlighting the need for context-specific solutions. Alamoush et al. (2022) further underline the significance of implementation schemes by port and public authorities, which are influenced by both contextual factors such as regulatory frameworks and operational factors including technological capabilities. These studies illustrate that contextual and operational considerations are essential for developing accurate evaluations and fostering sustainable investment in the maritime sector. In light of this, it is essential to assess the relationship between port characteristics and the adoption of GHG reduction measures, to better understand the enablers and the drivers associated with reducing the GHG emissions for a port. The current paper aims to fill this research gap by examining the characteristics influencing GHG reduction measures in European cargo and container ports and their implications for sustainable investing.

By focusing on the busiest ports, our study targets the primary contributors to GHG emissions within the maritime sector. This approach ensures that the analysis is both relevant and impactful, as it addresses the most significant sources of emissions and their corresponding mitigation efforts. The data available for these high-traffic ports are accessible and detailed, enabling a more comprehensive and precise analysis. That practical consideration enhanced the reliability and validity of our results, underscoring the importance of contextual and operational factors in evaluating sustainability efforts.

By identifying the port characteristics that are associated with sustainable investing in green measures, this study contributes to the broader literature on understanding the drivers and barriers to a green port development. Furthermore, these findings will be useful for future research as a starting point to explaining the correlation between port characteristics and sustainable investing in GHG reduction measures.

In the next section, we review the scientific literature on the barriers and drivers of sustainable investment in green ports, to build our rationale. Following this, Section 3 presents our variables, sampling, data collection, and method of analysis. This is followed by Section 4, with a visual representation of the distribution in the adoption of GHG reduction measures and descriptive statistics. In Section 5, we identify the port characteristics that might influence sustainable investment in GHG reduction measures by cargo and container ports in the European Union. The paper concludes with Section 6, which highlights the limitations and future research directions.

2. Literature Review and Hypotheses

2.1. Port Characteristics

European cargo and container ports are not identical to one to another; they differ vastly by their characteristics. These differences can be categorized into several key aspects. Firstly, ports differ by their location characteristics, including the nature of surrounding land (such as agricultural land, protected areas, forestry, open water, industry, or urban areas) (DNV-GL 2021), the type of access to the sea (marine inlet, engineered coastline, embayment, estuary, or river) (Puig et al. 2015; DNV-GL 2021; ESPO 2021), and the density of the population in the surrounding area (DeSombre et al. 2023; Sornn-Friese et al. 2021). For example, the port of Barcelona is situated on an engineered coastline in the city of Barcelona characterized by a high population density of 15,000 people per square kilometer. In contrast, the port of Nantes-Saint Nazaire is located in an estuary near the protected area of the Brière regional natural park.

Secondly, ports vary in size (Puig et al. 2015; DNV-GL 2021; ESPO 2021), which encompasses factors such as the tonnage of goods handled, the area of the port’s land, and navigable water. Moreover, ports differ by their sector of activity, which may involve a specialization in container handling, liquid bulk, dry bulk, roll-on/roll-off (RoRo) services, or a combination of these (DNV-GL 2021; ESPO 2021). For example, the port of Amsterdam, in 2019, handled 104 million tons of goods composed of 91% bulk, whereas the port of Valencia in the same year handled 65 MT of goods composed of 76% containers.

Finally, ports display variations in their governance structures, including different ownership models (public ownership, mixed public–private ownership, or private ownership)—where public ownership may differ by the level of public authority (state, municipality, province)—and types of organizations (mission-driven entity, non-economic public body, a profit-maximizing business) (ESPO 2022a, 2022b). Notably, only a small fraction (1%) of ports located in the European Union are privately owned, while the majority (77%) are mission-driven entities (ESPO 2022a, 2022b).

2.2. Barriers and Drivers of Green Port Sustainable Investing

The adoption of green measures by a port depends on the socio-political context of the country or the region in which the port is installed (Lawer et al. 2019; DeSombre et al. 2023). European ports are more focused on implementing climate change mitigation measures than West African ports and US ports (ESPO 2022a, 2022b; Lawer et al. 2019; DeSombre et al. 2023). West African ports have limited financial resources and face relatively little public pressure regarding climate change mitigation; therefore, they focus more on mitigating environmental issues such as improper waste management (Lawer et al. 2019). Green investments are influenced by the economic growth of the country (Eyraud et al. 2013) and highly depend on the public policies (Eyraud et al. 2013). European Union ports face top-down legal pressure, whereas green regulations are adopted at a supra-national level and focus on broader approaches (DeSombre et al. 2023). The EU has set the objective to be the first carbon-neutral continent by 2050, to mitigate the effect of climate change, and EU ports will have to comply with this objective. On the other hand, US ports face bottom-up legal pressure, where green measures are adopted by local governments focusing on more local issues such as improving the air quality of a port’s surrounding area (DeSombre et al. 2023). We supposed in this study that European Union countries are submitting to the same legal framework as one another.

The adoption of the Clean Air Action Plan (CAAP) in the ports of Los Angeles and Long Beach was a response to the social pressure of the local community (Giuliano and Linder 2013; Linder 2018). In this regard, not only are ports facing the concerns of the surrounding population but also of the final consumers of the goods transported in containers, who are more and more environmentally aware (Lam and Li 2019).

The adoption of GHG reduction measures by ports also depends on the type of shipping traffic (Styhre et al. 2017). Ports can easily direct their GHG reduction measures toward the ships calling by frequently.

Also, the port size matters; for instance, large ports communicate more than small ports about their environmental performance (Santos et al. 2016), and they are also more concerned about climate change (Puig et al. 2015).

In sum, the drivers and barriers to green ports can be linked to different port characteristics, as shown in

Table 1. Links between barriers and drivers of green ports and port characteristics.

Driver/Barrier	Port Characteristics	References
Financial capabilities of the port	Port size, economic growth of the country of location	(Ziegler and Seijas Nogareda 2009; Lawer et al. 2019; DeSombre et al. 2023; Eyraud et al. 2013; Vachon 2010)
Regulatory framework	Country of location, location near protected area	(DeSombre et al. 2023; Schoukens 2017; European Commission 2012)
Pressure from the local communities	Density of population	(Giuliano and Linder 2013; Linder 2018)
Pressure from the final customer	Type of activity (cargo mix)	(Lam and Li 2019)
Pressure from business partners	Type of activity (cargo mix)	(Lam and Li 2019)
Return on investment	Call frequency of ships, type of cargo handled	(Styhre et al. 2017)
Environmental priorities	Country of location, location in an estuary	(DeSombre et al. 2023; Schoukens 2017; European Commission 2012; ESPO 2021; Lawer et al. 2019)

.

2.3. Solutions for Ports to Reduce Their GHG Emissions

The primary step for ports to reduce their GHG emissions is to monitor and measure GHG (Pavlic et al. 2014; Lawer et al. 2019; ESPO 2021). Firstly, port authorities will be able to identify areas of operations that emit the most GHG and on which they need to focus to reduce their carbon footprint. Secondly, port authorities will be able to measure the efficacy of the solutions they are implementing.

Port GHG emissions can be classified into three scopes (WPCI 2010; The Greenhouse Gas Protocol 2004), as shown in

Table 2. Definition of a port’s scope 1, 2, and 3 emissions.

Scope	Definition
Scope 1	Port direct GHG emissions from activities and operations controlled by the port administration entity, such as port-owned fleet vehicles, port-administration-owned or -leased vehicles, buildings, and port-owned and -operated cargo handling equipment.
Scope 2	Port indirect GHG emissions from the generation of energy purchased by the port for its own buildings and operations.
Scope 3	Port indirect GHG emissions that are generally linked to the operations of tenants, including ships, trucks, cargo handling equipment, rail locomotives, harbor craft, tenant buildings, tenant-purchased electricity, and the commuting of both port and tenant employees.

.

The most significant source of overall GHG emissions by ports is maritime transport, that is, the vessels navigating toward and staying in a port (ESPO 2021; Gibbs et al. 2014). In the port of Rotterdam, the handling of containers and other goods accounts for only 2% of the overall CO₂ emissions of the port compared to 87% from the vessels calling at the port and 9% from the hinterland transport (Lechtenböhmer et al. 2018).

Port authorities cannot force port stakeholders to become greener by means of regulation (ESPO 2021). However, they can invest in infrastructure and provide financial incentives to encourage the decarbonization of the port stakeholders.

Vessel emissions in ports can principally be addressed by implementing on-shore power supply facilities (OSP) (Winkel et al. 2016), LNG bunkering stations (Peng et al. 2021; Balcombe et al. 2019), and vessel speed reduction (Linder 2018; Lindstad et al. 2011). Winkel et al. (2016) estimated that if all ports in Europe used cold ironing, EUR 2.94 billion of health costs could be saved as well as a potential reduction in carbon emissions of 800,000 tons in 2020. However, the benefit of cold ironing depends on the energy mix of the electricity supply (Winkel et al. 2016; ESPO 2021). Vessels using LNG emit 8–20% lower GHG emissions compared to those using distillate or residual fuel (Balcombe et al. 2019). A vessel speed reduction program can reduce the CO₂ emissions from shipping by 28% at zero abatement cost (Lindstad et al. 2011) and can be implemented straight away as it only consists of limiting the vessel speed.

Some researchers assert that greening hinterland transport is partly the port’s responsibility (Gibbs et al. 2014; Gonzalez-Aregall et al. 2018). Ports can promote a modal shift to rail and water transportation by constructing new infrastructure (intermodal terminal, railway) and implementing new technologies to make intermodal transport more attractive.

The energy consumption of port equipment can take the form of electricity or fuel (fossil fuels or clean fuels) (Iris and Lam 2019). Energy can be obtained from the grid in the form of electricity or it can be generated within the port. The GHG emission intensity of the generation of electricity purchased will depend on the energy mix of the country in which the port is implemented. The port can reduce its overall GHG emissions by producing its own electricity from renewable sources.

There are multiple technological options accessible for improving the energy efficiency and decreasing the greenhouse gas emissions of port operations (Dimitrov and Saraceni 2023). These options include utilizing electricity as an energy source, implementing autonomous vehicles (Saraceni et al. 2022), utilizing energy storage devices, adopting reefer cooling technologies, utilizing renewable energy sources, and employing cleaner fuels (Iris and Lam 2019). The electrification of RTG (rubber-tired gantries) could decrease GHG emissions by 67% compared to diesel-fueled conventional RTG (Na et al. 2017). The use of LNG as a fuel for port equipment reduces CO₂ emissions by 25% compared to fossil fuels (Na et al. 2017).

2.4. Hypotheses

Authorities of large ports are more concerned with reducing their GHG emissions than those of small ports (Santos et al. 2016; Puig et al. 2015). Furthermore, OSP and LNG bunkering facilities require high infrastructural investment from the port. Larger ports have more financial resources than small ports to invest in such infrastructure.

Therefore, the larger the port, the more it will implement GHG reduction measures.

H1. 

The port size is positively correlated with the adoption of GHG reduction measures by a port.

Styhre et al. (2017) suggest that there is a positive relationship between how often ships visit a port and sustainable investing in green measures to reduce GHG emissions by the port and the ship. Ships using alternative fuels such as LNG must guarantee that the fuel is available at the port visited. OSP requires a high capital investment from both ports and ships, and ship owners will have a return on investment only if the technology is used regularly. Container cargo ships have a higher rate of calling at port as they operate in a liner service that involves visiting the same ports on fixed rotation schedules. Meanwhile, some bulk cargo ships operate on a tramp service, without fixed routing and itineraries. where the route might be determined by market forces.

Container carriers are more concerned than bulk carriers with reducing their environmental impacts, mainly focusing on reducing their CO₂ emissions (Hegger and Saraceni 2024) (Lam and Li 2019). Consumers are increasingly placing the environmental impact as the criterion of choice when deciding whether or not to buy a product, and containers often transport consumer goods (finished or semi-finished products). However, consumer do not make a direct link between bulk shipping and the goods they consume.

Therefore, it is expected that European ports handling a high proportion of cargo containers implement more green solutions to reduce their GHG emissions than ports handling a high proportion of bulk carriers.

H2. 

The proportion of containers in the cargo mix handled by a European port is positively correlated with the adoption of GHG reduction measures.

Estuaries are highly productive natural habitats as they are a transition zone between the fresh water of the river and the saline water of the sea that both provides high levels of nutrients for the fauna and the flora. In addition to their goal of achieving carbon neutrality, ports situated in estuaries have a legal obligation to invest in the preservation and conservation of estuarine habitats (Schoukens 2017; European Commission 2012). Consequently, ports located in estuaries are less likely to have a priority concern related to the air (Puig et al. 2015) because they instead focus on the environmental protection and conservation of the area where they operate (Puig et al. 2015).

Therefore, ports located in estuaries are expected to implement fewer GHG reduction measures compared to ports located at sea as they also need to invest in the conservation of the natural habitat.

H3. 

Ports located in estuaries will implement fewer GHG reduction measures than ports located directly on the seafront.

Ports through measures like OSP, reducing vessels speeds, making alternative fuels available, and implementing financial incentives for ships to be more energy efficient have a direct impact in terms of lowering fuel emissions by ships in the seaport area. By using less fuels, ships will reduce their nitrogen oxide (NO_x), sulfur dioxide (SO₂), and particle matter (PM_2.5, and PM₁₀) emissions, which all pose risks to health of the people living in the port area (Saxe and Larsen 2004; Tzannatos 2010; Chatzinikolaou et al. 2015; Maragkogianni and Papaefthimiou 2015; Sorte et al. 2020).

Surrounding communities of the port area are concerned about the air quality and put pressure on the port to reduce the emissions generated by its activities (Sornn-Friese et al. 2021; Giuliano and Linder 2013; Linder 2018).

Therefore, ports located near densely populated areas are expected to implement more measures to reduce their GHG emissions to improve the air quality in the surrounding area due to higher pressure from the nearby population.

H4. 

The density of the surrounding population of the port is positively correlated with the adoption of GHG reduction measures by a port.

The economic wealth of a country is positively related to sustainable development within the country (Vachon 2010), and a higher level of income leads to higher investment in green technologies (Eyraud et al. 2013).

Therefore, ports located in countries with a high GDP per capita are expected to invest in more GHG emission reduction solutions than those located in countries with a lower GDP per capita

H5. 

A port located in a country with a higher GDP per capita will implement more GHG reduction measures that one located in a country with a lower GDP per capita

2.5. Theoretical Framework

These five hypotheses formulated lead to the theoretical framework illustrated in Figure 1. Our theoretical framework incorporates five port characteristics: port size, proportion of containers in the cargo mix, density of the surrounding population, access to the sea, and the country’s GDP for the port location. Each one these characteristics may have a positive or negative impact on the number of different GHG reduction measures adopted by ports, as indicated by the “+” or “-” signs in Figure 1.

3. Methodology

3.1. Predictive Variables

The selection of the five predictive variables in our theoretical framework (Figure 1) was based on a comprehensive literature review that we conducted in two phases. We first identified by which characteristics the ports differ from each other, by searching for different port organizations’ reports such as the ESPO (European Sea Port Organisation) and the IAPH (International Association of Ports and Harbors), and by using Google Scholar as a search engine to find peer-reviewed articles using the terms “port characteristics” and “port differences”. We selected port characteristics that could be objectively measured or quantified to facilitate the analysis. These included several port characteristics: size, access to the sea, density of surrounding population, country location, governance, types of industries located near the port, call frequency of ships visiting the port.

Secondly, we examined studies that investigated a relationship between port characteristics and sustainable investing in green measures by ports, or established drivers and barriers to green ports that could be linked to the port characteristics identified, to keep only the characteristics that might have an influence on the adoption of GHG reduction measures by ports. The results of this phase are presented in Table 1. Then, we kept only the characteristics that differed across the European Union ports under study. For this reason, we excluded the type of governance, as 99% of ports located in the EU are publicly owned (ESPO 2021). Meanwhile, the call frequency of the ships visiting the ports was not included in our theoretical framework because of the unavailability of public data.

The port size, proportion of containers in the cargo mix, density of the surrounding population, access to the sea, and economic wealth of the country location emerged as key predictive variables based on the literature review and their alignment with the criteria mentioned above.

3.2. Sampling

We used a non-probability sampling method, selecting the 33 busiest European Union container and cargo ports, as the data of these ports are accessible. Appendix A displays the list of European Union ports under study. Two external data sources were used to determine the busiest European port: the World Port Rankings from the American Association of Port Authorities in 2019; and the gross weight of goods handled in main ports by direction and type of cargo, using data from Eurostat. The 33 container and cargo ports under study handle 55% of the total container and cargo traffic in the European Union. By concentrating on the busiest ports, our study targeted the most significant contributors to GHG emissions in the maritime sector. This approach ensured that the analysis was relevant and impactful, addressing the primary sources of emissions and their mitigation efforts. The data for the busiest ports are accessible and detailed, which allowed for a thorough and accurate analysis. This practical consideration ensured the reliability of our results.

3.3. Data Collection

To measure the variables of our theoretical framework, we used secondary data sources. To assess the adoption of green measures for reducing greenhouse gas (GHG) emissions in ports, the total number of different green measures implemented by each port was measured. A port needs to adopt a combination of the different green measures available to achieve carbon neutrality. Yet, not all European ports disclose the full extent of their adopted green measures on their website or in their annual report (ESPO 2021). Thus, we relied on the dataset compiled by DeSombre et al. (2023), divided in two categories: port green measures that focus on scope 1 and 2 emissions, and ship green measures that focus on scope 3 emissions. The list of measures is provided in Appendix B. We summed these two categories to obtain the total number of green measures adopted by each port. This dataset was constructed by collecting information from port websites, annual reports, sustainability reports, and interviews with port officials. We verified the accuracy of the dataset by verifying the data of the Havre port and Rotterdam port.

We used satellite images from Google Earth to determine the port access to the sea, by assessing for each port if it is located on the seafront or in an estuary, which is the transition zone between river environments and maritime environments, as shown in Figure 2.

We used the gross weight of goods handled in each port to measure the port size. The mix of cargo handled was measured by computing the percentage gross weight of goods handled in containers in each port. For the port size variable and the percentage of containers in the cargo mix, we used the gross weight of goods handled in main ports by direction and type of cargo—quarterly data [MAR_GO_QMC] from Eurostat. The data were collected “by the national competent authorities in the reporting countries using a variety of data sources, such as port administration systems, national maritime databases, customs databases or questionnaires to ports or shipping agents” (Eurostat 2022).

We used the data from the full year 2019, to avoid the effects of COVID-19 on the gross weights of goods handled in 2020, 2021, and 2022.

To measure the population living around the port, we assessed the number of inhabitants within a 15 km radius of the port. With that purpose, we used the Gridded Population of the World fourth version of the Center for International Earth Science Information Network from Columbia University, which provides a spatially disaggregated population layer, as shown in Appendix C (Socioeconomic Data and Application Center (SEDAC) Columbia University. Measurement of Population Living within 15 km of the Port (example of Barcelona Port) (https://sedac.ciesin.columbia.edu/mapping/popest/gpw-v4/) accessed on April 2023.). The population data were from 2015. The 4-year difference between the population data and the data of other variables might have slightly biased the results, as the population does not grow at the same rate in each European Union area.

To measure the economic wealth of the country, we used 2019 data of the per capita GDP based on purchasing power parity from the World Bank. We excluded the GDP per capita of Ireland from the analysis because of the unique economic characteristics of Ireland. Since 2016, the GDP of Ireland has been rapidly growing because of favorable tax policies attracting multinational corporations to locate intellectual property assets in Ireland, causing a large increase in the country’s GDP without an actual increase in economic activity and productivity. As a result, Ireland’s GDP figures may not accurately reflect the true economic state of the country. By excluding Ireland’s GDP per capita, we aimed to enhance the accuracy and validity of the findings on the relationship between the GDP per capita and the adoption of green measures in ports.

Table 3. Data collection.

Variable	Measurement	Data Source	Year of the Data
Size	Annual tonnage of goods handled	Eurostat	2019
Access to the sea	Sea Estuary	Google Earth
Surrounding population	Number of people living within 15 km of the port	https://sedac.ciesin.columbia.edu/mapping/popest/gpw-v4/ accessed on 1 May 2023	2015
Mix of cargo handled	Percentage of containers	Eurostat	2019
Country’s wealth	Per capita GDP (PPP)	World Bank	2019
Adoption of GHG reduction measures	Amount of different GHG reduction measures implemented	DeSombre et al. (2023)	2020

summarizes the data collection.

3.4. Method of Analysis

We used a linear regression analysis to test the hypothesis of our theoretical framework. Linear regression is a statistical technique used to assess the association between a dependent variable and one or more independent variables. By using linear regression, we quantified the strength and direction of the relationships, determining if the relationships were significant and estimating the magnitude of the effect of each independent variable on the adoption of green measures. The linear regressions were run using the plm library in R.

Figure 3 shows the methodological chart.

4. Results

4.1. Descriptive Statistics

The 33 busiest European cargo and container ports have adopted on average seven different green measures to reduce their GHG emissions, with a standard deviation of seven. This reveals that the European ports are not moving at the same pace of implementing GHG emissions solutions to become carbon neutral. Figure 4 provides a visual representation of the distribution in the adoption of GHG reduction measures by the busiest European ports. Notably, five ports—Gothenburg, Hamburg, Marseille, Rotterdam, and Tallin—have adopted more than 13 different measures, indicating their strong commitment to becoming carbon neutral. On the other hand, six ports—Constanta, Gioa Tauro, Le Havre, Leixoes, Taranto, Venspils—have adopted four or fewer different measures, indicating the potential for improvement in their GHG reduction efforts.

Table 4. Descriptive statistics.

Variable	Mean	Median	Stdev	Min	Max
GHG reduction measures adopted	7.27	7	3.19	2	15
Port green measures (Scope 1 & 2)	5.45	5.00	2.54	1	11
Ship green measures (Scope 3)	1.82	2.00	0.94	1	4
Size	59,122.15	38,890	77,589.05	12,123	439,631
% of container in the cargo mix	0.28	0.22	0.24	0	0.86
Population	587,503.36	331,978	591,504.67	23,241	2,934,267
GDP Capita	46,149.83	45,799	8145.08	32,894	60,208

showcases the descriptive statistics.

On average, 587,503 people live within 15 km of the 33 busiest European ports, with a standard deviation of 600,676 inhabitants. Figure 5 shows that the density distribution of the population living within a 15 km radius of the ports is a right skewed, i.e., it is a lognormal distribution. Therefore, we used a logarithmic scale to analyze the correlation between the population variable and the other variables. This transformation allows for a better understanding of the relationship between population density and the adoption of green measures.

In terms of port size, four cargo and containers ports—Rotterdam, Antwerp, Hamburg, and Amsterdam—handle 25% of the European Union traffic, which represented 44.8% of the containers and cargo handled by the sample. On average, the ports in our sample handled 59,122 tons of goods in 2019, with a standard deviation of 78,792. Figure 6 shows that the density distribution of the port size is right skewed. Therefore, we used a logarithmic scale to analyze the correlation between the size variable and the other variables.

The cargo mix of the 33 European Union container and cargo ports under study is 33% containers, 37% liquid bulk, and 18% dry bulk, with a standard deviation of 24% for the proportion of containers in the cargo mix. This diverse cargo composition reflects the different specializations and trade patterns observed among the ports.

Overall, 45% of the ports under study are located in estuaries and the other 55% are located on the seafront.

The average GDP per capita of the countries in which the ports of our sample are located is USD 46,150 per year, with a standard deviation of USD 8275 per year.

The descriptive statistics demonstrate the heterogeneity of the characteristics of the European ports under study. They differ vastly in term of their access to the sea, density of surrounding populations, cargo mix, and size. This diversity highlights the varied contexts in which these ports operate.

4.2. Correlation between Independent Variables

The correlation table (

Table 5. Correlation table.

	log(Size)	GDP per Capita	log(Population)	% Container
log(Size)	1
GDP per Capita	0.4 *	1
log(Population)	0.49 **	0.16	1
Container	0.26	−0.03	0.17	1

** p-value < 0.001, * p-value < 0.05.

) shows that the population variable and the port size variable are significantly positively correlated, with a correlation coefficient of 0.49 in our sample. This suggests that there is a moderate degree of association between the size of the port and the population density in the surrounding area. The port size is also significantly correlated with the GDP per capita. This finding suggests that the bigger ports are located in the richest countries of the EU.

4.3. Linear Models

The regression analysis (

Table 6. Single linear regression model—total GHG reduction measures adopted.

	Xi1	β0	β1	Adjusted R-Squared
Model 1	GDPP per Capita	0.03	0.00016 *	0.13
Model 2	Log(Size)	−14.7 *	2.07 **	0.22
Model 3	Log(Population)	−9.29	1.3 **	0.18
Model 4	Sea Location	7.8 ***	−0.97	−0.009
Model 5	% of container	7.18 ***	0.31	−0.031

Model: yi = β0 + β1xi1; yi = number of GHG reduction measures adopted; ** p-value < 0.001, * p-value < 0.05. *** all p-values less than 0.001 are summarized with three asterisks.

) revealed a significant positive relationship between port size and the number of green measures adopted by ports to mitigate their GHG emissions. The statistical analysis yielded a p-value of 0.3%, indicating a strong level of significance. This result suggests that larger ports, with a higher volume of container and cargo handling, tend to implement a greater number of green measures to mitigate their GHG emissions (Figure 7). The adjusted R-squared value of 0.22 suggests that approximately 22% of the variation in the number of different green measures implemented by European Union ports could be explained by the size of a port. Notably, the influence of port size was observed across different categories of green measures, including those targeting scope 1 and 2 emissions (p-value = 0.0075) as well as those addressing scope 3 emissions (p-value = 0.0081) (

Table 7. Single linear regression model—scope 1 and 2 reduction measures adopted.

	Xi1	β0	β1	Adjusted R-Squared
Model 6	GDPP per Capita	0.94	0.000097.	0.09
Model 7	Log(Size)	−10.64	1.52 **	0.18
Model 8	Log(Population)	−7.67	1.03 *	0.18
Model 9	Sea Location	5.8 ***	−0.7	−0.016
Model 10	% of container	5.4 ***	0.18	−0.031

Model yi = β0 + β1xi1; yi = number of scope 1 and 2 reduction measures adopted; ** p-value < 0.001, * p-value < 0.05. *** all p-values less than 0.001 are summarized with three asterisks.

and

Table 8. Single linear regression model—scope 3 reduction measures adopted.

	Xi1	β0	β1	Adjusted R-Squared
Model 11	GDPP per Capita	0.91	0.00006 **	0.24
Model 12	Log(Size)	−4.07.	0.55 **	0.18
Model 13	Log(Population)	−1.61	0.27	0.076
Model 14	Sea Location	2 ***	−0.33	−0.0002
Model 15	% of container	1.8 ***	0.12	−0.031

Model yi = β0 + β1xi1; yi = number of scope 3 reduction measures adopted; ** p-value < 0.001. *** all p-values less than 0.001 are summarized with three asterisks.

). This finding supports the first hypothesis that port size impacts the adoption of green measures.

The regression analysis showed a significant positive impact of the population living withing 15 km of the port on the number of different green measures adopted by the port to mitigate its GHG emissions. The regression analysis yielded a p-value of 0.7%, which indicates a strong level of significance (Table 6). However, the density of the population only significantly impacts the number of green measures targeting a reduction in the port’s scope 1 and scope 2 emissions (p-value = 0.007) (Table 7), while it did not demonstrate a significant impact on the number of green measures focusing on scope 3 emissions at the conventional significance level of 5% (p-value = 0.067) (Table 8). This finding implies that ports located in areas with a higher population density tend to adopt a greater number of green measures specifically targeting a reduction in their operational emissions (scope 1 and 2). These results provide support for the fourth hypothesis of our study, indicating a positive influence of population density on the adoption of green measures by ports (Figure 8).

Moreover, the results of the linear regression analysis reveal a significant relationship between the GDP per capita of the country in which the port is located and the number of different green measures adopted by the port to mitigate its GHG emissions (Table 6). The p-value associated with the GDP per capita variable is 2.8%, indicating a statistically significant relationship. However, the adjusted R-squared value of 12.6% is relatively low, indicating that the GDP per capita variable explains only a small portion of the variation in the total number of green measures implemented by the port. Notably, when examining the adoption of green measures specifically targeting scope 3 emissions, we found that the GDP per capita of the port’s country exhibited a significantly positive relationship (p-value = 0.0024), with an R-squared value of 24%, indicating that it explains a large proportion of the variation in the number of scope 3 green measures adopted by the port (Table 8). Conversely, the GDP per capita variable did not demonstrate a significant impact on the adoption of measures solely targeting scope 1 and 2 emissions (p-value = 0.08) at the conventional significance level of 5% (Table 7). These results provide support for the fifth hypothesis of our study, indicating the positive influence of the GDP per capita of the country in which the port is located on the adoption of green measures by ports.

The single linear regression analysis conducted indicated that the access to the sea of the port (sea or estuary) and the cargo mix handled by the port do not have a statistically significant impact on the number of different green measures adopted by the port to reduce its GHG emissions (Table 6). This suggests that the selection and implementation of green measures are not heavily influenced by these two factors. Therefore, the second and third hypotheses are rejected.

As the log(population) and the log(size) are positively correlated with correlation coefficient of 0.48 in our sample, and also log(population) with the GDP per capita, this suggests that part of the effect of the density of the population living around port on the number of different green measures adopted by the port might be attributed to the port size or vice versa. The effects of the port size and the GDP per capita on the number of green measures adopted might also be interlinked.

To further investigate the impacts of the size, population, and GDP per capita variables, we ran multiple linear regression models. These suggested that the potential positive effect the port size might have on the adoption of different green measures by European ports is interlinked with the effect of the GDP per capita and the surrounding population (see

Table 9. Multiple linear regression model—2 predictive variables.

	xi1	xi2	β0	β1	β2	Adjusted R-Squared	Model p-Value
Model 16	GDPP per Capita	Log(Size)	−15.27 *	0.00009	1.72 *	0.25	0.006
Model 17	GDPP per Capita	Log(Population)	−13.63 *	0.00013 *	1.17 *	0.27	0.004
Model 18	Log(Size)	Log(Population)	−18.82 *	1.5 *	0.8	0.26	0.004

Model: yi = β0 + β1xi1 + β2xi2, yi = number of GHG reduction measures adopted. * p-value < 0.05.

and

Table 10. Multiple linear regression model—3 predictive variables.

	xi1	xi2	Xi3	β0	β1	β2	β3	Adjusted R-Squared	Model p-Value
Model 19	GDPP per Capita	Log(Size)	Log(Population)	−19.13 *	0.0001	1.08	0.82	0.29	0.006

Model: yi = β0 + β1xi1 + β2xi2 + β3xi3, yi = number of GHG reduction measures adopted. * p-value < 0.05.

). The adjusted R-squared does not vary much between model 2 and models 16 and 18. Furthermore, the beta values of the GDP per capita and log(population) become insignificant when they are in the same model as log(size); however, they are significant when put together without log(size).

5. Discussion

The purpose of this research was to determine what port characteristics might influence sustainable investing in GHG reduction measures by cargo and container ports in the European Union. Data were collected on the 33 largest European Union ports across several secondary databases. The findings suggest there are three significant port characteristics related to sustainable investing in different GHG reduction measures by European Union cargo and container ports: port size, the density of the population living around the port area, and the GDP per capita of the country of location. The findings suggest that port size is the most correlated characteristic with sustainable investing in GHG reduction measures by ports. Two port characteristics—the mix of cargo handled and the access to the sea of the port—have no influence on sustainable investing in GHG emissions by ports in the European Union. Figure 9 summarizes the results of our investigations.

5.1. Port Size

In our first hypothesis, we proposed that port size is positively correlated with sustainable investing in GHG emissions reduction measures by ports. Our results support this first hypothesis, suggesting that larger ports, which handle a higher volume of container and cargo traffic, are more inclined to adopt a greater number of green measures to mitigate their GHG emissions. Larger ports might have greater financial capabilities and technical capabilities to invest in and implement GHG emissions reduction measures. Our finding is consistent with Puig et al. (2015) and Santos et al. (2016), who found that larger port authorities in the European Union display greater concern for reducing emissions compared to smaller port authorities. Since 2015, it seems this trend has not changed; smaller European Union ports are still putting less effort into becoming carbon neutral than larger ports. However, as the port size is correlated with the GDP per capita of the country location and the density of the surrounding population, the causal effect on sustainable investing in green measures by ports might be interlinked with these other two factors.

5.2. Cargo Mix

In our second hypothesis, we proposed that the proportion of containers in the cargo mix handled by European ports is positively correlated with sustainable investing in GHG reduction measures. However, our results lead us to reject this hypothesis. Part of the hypothesis is built upon the assumption that container cargo has a higher call rate at ports as the majority is operated in a liner service, unlike bulk cargo, which can be operated in a tramp service. However, we did not find any publicly available database on the proportion of bulk cargo operated in tramp services visiting European Union ports that would allow us to verify this assumption. Our findings contrast with Lam and Li’s (2019) findings, which highlight that container carriers are more concerned than bulk carriers with reducing their environmental impact. Our study suggests that there is no significant difference between the pressures exerted by container carriers and bulk carriers on ports to adopt greener practices.

5.3. Estuary Location

In the third hypothesis of our study, we proposed that ports located in estuaries implement fewer GHG reduction measures than those located directly on the seafront. Our results do not support this hypothesis. Contrary to the findings of Puig et al. (2015), which suggested that estuary ports prioritized the preservation of estuarine habitats over air emission measures, our study did not find a significant impact of the water location on sustainable investing in different GHG reduction measures by European Union ports. This indicates a shift in trends since 2015. At that time, climate change was not a top 10 environmental priority of European Union ports; it since became, in 2022, the top priority of the majority of European Union ports (ESPO 2021), irrespective of their water location. Our findings indicate that ports located in estuaries, despite the need to preserve estuarine habitats (Schoukens 2017; European Commission 2012), exert similar efforts as ports situated on the seafront to reduce their GHG emissions. Therefore, it appears that the estuary location of ports may no longer be a determining factor in sustainable investing in GHG reduction measures.

5.4. Density of the Surrounding Population

In our fourth hypothesis, we proposed that the density of the surrounding population of the port is positively correlated with sustainable investing in GHG reduction measures by the port. Our results support this hypothesis. These findings align with previous research by Sornn-Friese et al. (2021), Giuliano and Linder (2013), and Linder (2018), who highlighted the role of the surrounding population in pressuring ports to mitigate their GHG emissions. As the population density around the port increases, the pressure on the port to adopt green measures intensifies, driven by the surrounding communities and local political authorities. However, our results revealed an unexpected finding: ports tend to prioritize the reduction of scope 1 and 2 emissions in response to this pressure, rather than addressing their scope 3 emissions. It is important to consider that scope 3 emissions from ports are significantly higher than scope 1 and 2 emissions, and these emissions, including particulate matter, directly impact human health.

5.5. Port Location Country’s Wealth

In the fifth hypothesis of our study, we proposed that a port located in country with a higher GDP per capita will implement more GHG reduction measures that the one located in country with a lower GDP per capita. Our results support this hypothesis. This can be attributed to several factors, such as the higher environmental awareness of people in wealthier countries and the greater financial capacity of ports in wealthier countries to invest in green measures. Our results align with previous research (Vachon 2010; Eyraud et al. 2013) that states that firms located in wealthier countries tend to invest more in green technologies and sustainable development. However, our study produced an unexpected result: the impact of the wealth of the country in which the port is located was observed only in relation to scope 3 emission reduction measures, with no impact on measures targeting emissions from port operations (scope 1 and 2). This suggests that ports in wealthier countries may place a stronger emphasis on addressing the broader environmental impact of their operations.

5.6. Implications

Just as ports are pressured to adopt measures to reduce their carbon footprint, investors are increasingly focusing on sustainability criteria when making investment decisions. Sustainable investing involves incorporating environmental, social, and governance factors into investment analysis and portfolio construction. By analyzing port characteristics that influence GHG reduction measures, we can develop a framework that investors might use to assess the sustainability practices of various companies or industries. The findings of this study have significant implications for the future sustainability of European Union ports. In regard to the obligation for all European Union ports to become carbon neutral by 2050, it is evident that concerted efforts will be required to achieve this ambitious goal. Ports will need to use a mix of several different green measures in order to effectively reduce their GHG emissions and mitigate their environmental impact. However, our study reveals that there is still a long way to go, particularly for smaller ports, those located in low-density populated areas, and those located in countries with a relatively smaller economies. It is imperative for authorities and policymakers to prioritize and allocate adequate attention, resources, and support to these ports. By focusing on the specific needs and challenges faced by smaller ports and those in low-density areas, authorities can facilitate and accelerate their transition toward carbon neutrality. This could be achieved through a combination of targeted incentives, financial support, capacity-building programs, and collaborative initiatives with local communities and stakeholders.

6. Conclusions

Becoming carbon neutral has become one of the primary objectives of the European Union, and yet it will be challenging for ports to sufficiently reduce their carbon emissions. Today, ports are moving at different paces toward carbon neutrality. This study contributes to the literature on the drivers and barriers that ports face to adopting and implementing GHG reduction measures by studying five potential characteristics—a port’s size, the density of the surrounding population, its access to the sea, the cargo mix, and the wealth of the country in which it is located. Among the five characteristics under study, we identified three potential drivers of sustainable investing in GHG reduction measures by ports in the EU—port size, density of the population, and GDP of the country of the location. These findings are in line with the previous literature. However, the density of the surrounding population only influences the adoption of measures targeting the scope 1 and 2 emissions, and the GDP of the country of the location only influences the adoption of measures targeting the scope 3 emissions. Contrary to Puig et al.’s findings in 2015, we have found that the access to the sea of the port has no influence on the adoption of green measures; therefore, it seems some barriers and drivers have changed throughout the years. We have also found that the cargo mix handled by a port has no significant impact on the adoption of GHG reduction measures by ports in the European Union.

These finds have implications and offer critical insights for sustainable investing. Larger ports, supported by greater financial and technical resources, are more likely to implement green measures to reduce GHG emissions. This highlights an investment opportunity in larger ports for those focused on environmental sustainability. On the contrary, the cargo mix does not significantly influence sustainable practices, suggesting that investors should look beyond cargo types when assessing sustainability efforts. Ports situated in densely populated areas tend to adopt more rigorous GHG reduction measures, driven by community and political pressures. This indicates that sustainable investors might prioritize ports in such regions due to their proactive environmental strategies. Furthermore, ports in wealthier countries are more inclined to address broader environmental impacts, particularly scope 3 emissions, presenting another focal point for sustainability-focused investments. This study underscores the need for targeted support and resources for smaller ports, those in low-density areas, and those in economically less affluent countries. Policymakers and authorities must prioritize these ports to ensure a uniform transition toward carbon neutrality by 2050. For sustainable investors, this presents an opportunity to influence and accelerate environmental progress through strategic investments and collaborations that support under-resourced ports.

Ultimately, this research supports the development of a nuanced investment framework that accounts for the port size, location, surrounding population density, and national wealth, guiding investors toward more sustainable and impactful investment decisions in the maritime sector.

Limitations and Future Research

Several limitations should be considered when interpreting the findings of our study. The sample used in our study is composed of the busiest European Union ports, and the results may not be generalizable to smaller ports in the EU. Future studies should include a broader range of port sizes to capture what port characteristics are associated with the adoption of GHG reduction measures in smaller European Union ports.

Additionally, we focused on the adoption of different green measures, giving weight to each measure without considering the extent to which it is implemented or its specific impact on GHG emissions reduction. Therefore, we did not measure the actual effectiveness of the adopted measure by each port. Assessing the true progress made by ports in reducing their GHG emissions would require data from emissions monitoring and the historical trends. However, such data are currently limited, as only a fraction of European Union ports monitor their GHG emissions, and historical records are scarce, with only a few ports measuring their emissions in 2015 (ESPO 2021). We used data from the full year 2019, to avoid the effects of COVID-19 on the gross weights of goods handled in 2020, 2021, and 2022. Nonetheless, we contend that it is not feasible to draw general conclusions about GHG reduction measures based on one year of data. However, our aim was to identify port characteristics that might influence sustainable investments in GHG reduction measures by cargo and container ports. Future studies could focus specifically on fluctuations in reduction measures, using a longer period to draw general conclusions about GHG reduction measures.

Furthermore, in our study, we employed linear regression analysis to examine the significance of the relationships between the adoption of GHG reduction measures and port characteristics. While this approach provides insights into the statistical significance of the relationships, it does not establish causal relationships. Therefore, we can only make assumptions about the causality. Future research could explore the underlying mechanisms through which, in European Union ports, density and port size influence the adoption of green measures and how these can differ from port to port.

The non-probability sampling method has limitations, particularly in terms of generalizability to smaller ports or those with less traffic. Future research could expand the sample to include a broader range of ports, incorporating robustness tests and probabilistic sampling methods to enhance the robustness of the findings. The low R-squared values suggest that the chosen model might not explain the variance in the data well. However, the approach used in selecting the data for this study addresses potential concerns about the reliability of the results.

Non-probability sampling and data accessibility were prioritized in approach, whereby we deliberately selected the 33 busiest European Union container and cargo ports. By concentrating on the busiest ports, which handle 55% of the total container and cargo traffic in the European Union, we ensured this study’s relevance and impact. These ports are the primary contributors to greenhouse gas (GHG) emissions in the maritime sector. We addressed the most significant sources of emissions with the goal to make this study meaningful and impactful, thereby compensating for the low R-squared values. In addition, our study relied on established secondary data sources, such as the World Port Rankings and Eurostat data, which are reputable and comprehensive. This reliance on robust data sources supports the reliability of the findings despite the low explanatory power indicated by the R-squared values. We also ensured thorough data verification, with the accuracy of the dataset verified by cross-checking the data for specific ports (e.g., Le Havre and Rotterdam). This additional step of verification ensured the data used were accurate and reliable, thus mitigating concerns arising from low R-squared values.

Future research could investigate other contextual and situational factors such as the use of surrounding land (agricultural land, protected areas, forestry, open water, industry, city), support from local authorities, and environmental regulations of the country. Lastly, extending the geographical scope of the research to ports located in other continents or regions would enable a more comprehensive understanding of the influence of port characteristics on the adoption of green measures. Comparing and contrasting the findings from different geographical areas would provide valuable insights into the contextual nuances and regional variations in port sustainability practices.

By addressing these research gaps and considering a broader range of factors and geographical contexts, future studies can enhance our understanding of the complex dynamics influencing the adoption of green measures by ports.