Key criteria influencing the choice of Arctic shipping: a fuzzy analytic hierarchy process model

ABSTRACT

As Arctic sea ice shrinks due to global warming, the Northern Sea Route (NSR) and the Northwest Passage (NWP) offer a substantial reduction in shipping distance between Asia and the European and North American continents, respectively, when compared to conventional routes through the Suez and Panama Canals. However, Arctic shipping routes have many problems associated with their use. The main objective of this paper is to identify the key criteria that influence the decisions of shipping operators with respect to using Arctic shipping routes. A multi-criteria decision-making methodology, the Fuzzy Analytic Hierarchy Process, is applied to rank four potential categories of criteria (‘economic’, ‘technical’, ‘political’ and ‘safety’ factors) and their sub-criteria.

The results of the analysis suggest that, on aggregate, ‘economic’ is the most important category of influential factors, followed by ‘safety’, ‘technical’ and ‘political’ factors. The paper concludes, however, that the most influential specific sub-criteria relate to risks that lie mainly within the ‘safety’ and ‘political’ domains and that, especially in combination, these overwhelm the importance which is attached to ‘economic’ factors such as reduced fuel use. Finally, the implications of these findings for the future development of Arctic shipping are addressed at a strategic level.

KEYWORDS:

• Arctic shipping
• analytic hierarchy process
• route choice
• shipping companies
• shipping operators

1. Introduction

Projections derived from climate models suggest that one of the adverse consequences of global warming is that summer sea ice in the Arctic Basin will retreat further and further away from most Arctic landmasses. As a consequence of this, previously frozen areas in the Arctic may become seasonally or permanently navigable. The Northern Sea Route (NSR) and the Northwest Passage (NWP) have emerged, therefore, as increasingly realistic alternatives to more conventional long-distance sea routes utilizing the Suez and Panama Canals, despite the expansion of the latter (Wang, Talley, and Brooks 2016). For the foreseeable future, however, not only will these emerging routes probably not be navigable for more than 5 or 6 months a year, but there also exist many restrictions or barriers to their practical feasibility. For example, ship design for Arctic shipping requires additional investment in hull construction. In addition, ships that transit Arctic waters will also require specialized seafaring skills in both officers and crew and this will mean that manning costs are higher than normal. Harsh climatic conditions (such as floating ice sheets, icebergs, fog and violent winds) affect navigation and bring unpredictable risks. Obviously, this will have an impact on insurance costs (Berglund, Kotovirta, and Seina 2007).

For shipping companies, the raison d’être for choosing to transit the NSR or the NWP is to reduce both the distance and time involved in intercontinental shipping and, ultimately, therefore, to reduce both the direct and opportunity cost associated with moving freight through conventional routes (Beveridge et al. 2016). At the same time, it is clearly important to maintain a reliable shipping service for customers.

Pruyn (2016) points out that since 1995 there have been an emerging number of works investigating the feasibility of Arctic shipping routes. Most of these papers have focused on their economic viability, followed by safety and other aspects. However, there are many uncertainty factors affecting the transit time of Arctic shipping routes and different sources have adopted a number of different perspectives on this. There is, therefore, a continuing need to codify the different analyses and to effectively structure and categorize the potential key criteria in the decision to opt for an Arctic shipping route.

There exists, however, a range of technical, financial and operational factors that may affect both the cost and quality of the shipping service offered and, more specifically, the route choice decisions of shipping companies. Through a comprehensive literature review and the analysis of expert opinion, the objective of this paper is to investigate the various possible factors that affect a shipping company’s intention whether or not to use Arctic shipping routes. The paper is organized as follows: Section 2 reviews past studies related to Arctic shipping issues. Section 3 describes the Fuzzy Analytic Hierarchy Process (FAHP) methodology and applies it as the basis of the empirical analysis conducted within this work. Section 4 provides the results of this analysis, while Section 5 contains a discussion of these results and the conclusions which can be drawn.

2. Literature review

2.1. The geography of Arctic shipping

As shown in Figure 1, the Arctic shipping routes considered in this paper are the NSR and the NWP.¹ The NSR is sometimes referred to as the Northeast Passage and is a shipping route between the Atlantic Ocean and the Pacific Ocean which runs along the Russian coast of Siberia and the Far East. It includes five Arctic Seas: the Barents Sea, the Kara Sea, the Laptev Sea, the East Siberian Sea and the Chukchi Sea. The NWP encompasses a coastal route through the Canadian Archipelago and the North of Alaska.

Figure 1. The Northern Sea Route (NSR) and the Northwest Passage (NWP).

For many years, the NSR had been discussed as a potential new and viable route for commercial shipping between Europe and Asia. Most recently, Yumashev et al. (2017) indicated that the NSR offers shorter travel distances of up to 40% between Asia and Europe. In August 2009, however, two new ice-strengthened heavy-lift vessels of the German company Beluga transited the NSR from East-to-West as part of a small convoy escorted by a Russian nuclear ice-breaker. Russian ice pilots were onboard for part of the voyage. These were the first western commercial vessels to make this transit. The President of Beluga Shipping said the voyage saved each vessel about €300,000 compared to the normal Korea-to-Rotterdam route via the Suez Canal (Paterson 2009; Franckx and Boone 2012). The saving in distance in comparing the two alternative routes is graphically illustrated in Figure 2.

Figure 2. The distance saved using the Northern Sea Route (NSR) for a voyage from Rotterdam to Korea.

Irrespective of the obvious savings in both distance and time which both the NSR and the NWP provide, there still exist many factors that need to be considered before opting for either of these routes. For example, there are higher building and fuel costs for ice-class ships (Solakivi, Kiiski, and Ojala 2017), there is a need for slower and less consistent speeds, there are navigational difficulties and greater risks and also significant additional direct costs are incurred because ice-breaker services are required. Therefore, before deciding in favour of an Arctic shipping route for a specific shipping movement, the following factors should be considered: the specific route recommendation, communication services, manning issues (i.e. the experience of officers and crew), the availability of charts and maps, special vessel steerage, H&M insurance, P&I insurance, repairs and maintenance and the supply of fresh water and food (Liu and Kronbak 2010; Kiiski 2017).

2.2. Key criteria

It is clear that there exist a number of key criteria that need to be evaluated when a shipping company arrives at a decision whether or not to opt for an Arctic shipping route. Within this paper, these key criteria have been categorized into four main groups of key criteria. It is important to recognize, however, that these major categories of criteria could never be considered to be wholly discrete; there is inevitably significant potential for overlaps between them, not least because virtually all the sub-criteria which fall within a specific category of key criteria are likely to have economic implications through the imposition of additional costs. The corollary of this is that the categorization system and process may be perceived as potentially subjective and, therefore, contentious. However, the categorization system applied within this study is largely based on accepted precedent set in previous works and, in any case, the potential shortcomings associated with the categorization system are overcome by the analysis of the influence of individual sub-criteria. Bearing all this in mind, the four key criteria that comprise the categorization system can be described as follows:

2.2.1. Economic factors

These relate to the direct cost and time issues shipping companies need to consider when evaluating the use of an Arctic shipping route (Verny and Grigentin 2009; Somanathan, Flynn, and Szymanski 2009; Liu and Kronbak 2010; Schøyen and Bråthen 2011; Xu et al. 2011; Lasserre 2014; Pruyn 2016; Meng, Zhang, and Xu 2017). Generally, capital cost, operating costs (e.g. crew costs and insurance) and voyage costs (e.g. fuel costs and transit fees) are relevant when evaluating shipping costs (Meng, Zhang, and Xu 2017). Here, economic factors can be divided into fuel costs, crew costs, insurance costs, other transit costs and sailing time:

• Fuel costs includes bunkers, diesel and lubricating oil (Laulajainen 2009; Ho 2010; Brubaker and Ragner 2010; Xu et al. 2011., Schøyen and Bråthen 2011; Hong 2012; Furuichi and Ostuka 2013; Lasserre 2014; Bourbonnais and Lasserre 2015; Lindstad, Bright, and Strømman 2016). Fuel consumption depends on distance, speed and fuel efficiency, as well as other factors.

• Crew costs relate to both officers and other crew members (Liu and Kronbak 2010; Furuichi and Ostuka 2013; Lasserre 2014). The harsh polar climate requires an experienced captain and crew. This places a burden of higher salary costs on shipping companies.

• Insurance costs in relation to both H&M (hall and machinery) and P&I (protection and indemnity) are higher than would normally be the case due to greater than average risks, and higher potential for accidents, associated with transiting polar waters (Lasserre and Pelletier 2011; Xu et al. 2011; Furuichi and Ostuka 2013; Lasserre 2014). Pruyn (2016) goes so far as to suggest that, at the time of writing, insurance coverage is simply not available for ships transiting the NSR, since the chances of damage to the hull are so large.

• Other transit costs that are likely to be incurred relate to ice-breaking services, increased maintenance and depreciation of ships, administration costs, etc. (Johnston 2002; Xu et al. 2011; Furuichi and Ostuka 2013; Pruyn 2016).

• Sailing time may be longer than planned due to delays caused by poor visibility, difficult ice conditions, unforeseen repairs, etc., despite the fact that Arctic routes between Europe and the Far East are shorter than through the Suez Canal. Such delays may have an impact on any or all of a vessel’s transit time, average speed, time in port and at berth (Lasserre and Pelletier 2011; Xu et al. 2011; Pruyn 2016). Meng, Zhang, and Xu (2017) pointed out that sea, weather and ice conditions may increase sailing time uncertainties in transarctic shipping. Indeed, very severe ice conditions will increase routing difficulties and, therefore, also affect the travel distance and time (Meng, Zhang, and Xu 2017).

Again, these are all evaluated in comparison to the traditional route from the Far East to Europe which transits the South China Sea, the Straits of Malacca, the Indian Ocean, the Suez Canal, the Mediterranean Sea and the English Channel. The focus of such analyses is on the savings and/or additional costs incurred as a result of utilizing an Arctic shipping route as compared to the conventional alternative.

When evaluating the alternative routes, many papers focus on a comparison of the direct costs (Verny and Grigentin 2009; Liu and Kronbak 2010; Schøyen and Bråthen 2011; Liu 2016). The NSR provides an increasingly viable alternative route which, according to Liu and Kronbak (2010), Schøyen and Bråthen (2011) and Lindstad, Bright, and Strømman (2016), could save, in some cases, approximately 40% of the sailing distance (e.g. from Yokohama to Rotterdam). The economic advantage of the NWP is less certain, with Somanathan, Flynn, and Szymanski (2009) finding that, compared to using the Panama Canal, the required freight rate for a transit from St. Johns to Yokohama is only slightly lower using the NWP and is actually slightly higher for a transit from New York to Yokohama using the NWP.

Despite the consistent finding of past studies that there exists clear economic potential in transiting the Arctic route via the NSR, the marginal nature of the economic potential once all costs are evaluated, together with the challenging physical environment and associated safety factors, as well as the uncertain political climate, all conspire to constrain the volume of shipping activity in the Arctic area. Most crucially, therefore, it is very important to properly evaluate the less obvious (hidden) costs of Arctic transits (Brubaker and Ragner 2010; Schøyen and Bråthen 2011).

2.2.2. Technical factors

These refer to specific technological, ship design or port infrastructure issues that might affect Arctic shipping. As emphasized earlier, that is not to say that they do not also have economic consequences in terms of increasing costs. For the purposes of the categorization applied herein, technical factors can be divided into the need for ice-breaking services, ship construction, navigation and communication:

• Ships transiting Arctic routes should be escorted by an ice-breaker. As a result, the average speed is slower than along the conventional route (Xu et al. 2011).

• With regard to ship construction, an ice-class ship with a thicker hull, equipped with an appropriate propulsion system must be utilized (Johnston 2002; Laulajainen 2009; Lasserre and Pelletier 2011; Hong 2012).

• Advanced navigational aids and effective ice information (e.g. GPS/GLONASS—Global Navigation Satellite System) are necessary in order to provide precise and correct route recommendations (Brubaker and Ragner 2010; Liu and Kronbak 2010; Schøyen and Bråthen 2011; Farré et al. 2014).

• Real-time communication with detection radar is important for transiting Arctic waters (Ho 2010; Liu and Kronbak 2010; Farré et al. 2014). The lack of hydrographic data within Arctic waters was highlighted as a particular technical problem by Pruyn (2016). Similarly, there are still serious limitations regarding radio and satellite communications in the Arctic region due to poor satellite coverage (Meng, Zhang, and Xu 2017).

It is important to recognize that despite the reduction in summer ice cover in the north polar seas, navigation remains difficult due to complex environmental conditions such as the prevailing ice conditions, extreme cold, fog, storms, darkness, etc. Fundamentally, therefore, ice class ship designs with thicker hulls and greater structural support, in addition to pilotage, remain mandatory for vessels in this area.

2.2.3. Political factors

These relate to the political issues of governance, jurisdiction, regulation, legal problems and conflicts of interest among the stakeholders in the Arctic region. Political factors can be divided into the attitudes of coastal countries in the region, maritime conventions and stakeholder concerns:

• What are the attitudes of Russia and Canada towards the possible opening up and commercialization of Arctic shipping routes (Johnston 2002; Laulajainen 2009; Brubaker and Ragner 2010; Schøyen and Bråthen 2011; Hong 2012; Weber 2012)? Further, Arctic states, such as the USA, Norway and Denmark also defend their political interests in this region. In the case of the NSR, for example, most ships will need to be accompanied by a Russian ice-breaker to traverse the icy waters. Securing ice-breaking services at an appropriate price will involve complex political relations (Meng, Zhang, and Xu 2017).

• In order to make Arctic routes more amenable to international use, maritime conventions relating to vessel rights and obligations within these waters still need to be established (Pelletier and Lasserre 2012).

• Arctic resources (e.g. oil, gas, etc.) have prompted many non-Arctic countries (e.g. China, South Korea, Japan, Singapore, etc.) to voice their strategic interest in the region (Johnston 2002; Ho 2010; Huang, Lasserre, and Alexeeva 2014).

Although the concept of the ‘freedom of the seas’ is very germane to this arena, so too are the key jurisdictional roles played by Russia and Canada within the NSR and NWP, respectively (Laulajainen 2009; Ho 2010). Other major Arctic countries with a direct interest in shipping activity within the region are the United States, Greenland, and Norway. The neighbouring Nordic countries of Finland, Iceland and Sweden also have vested interests associated with the future easier exploitation of natural resources in the Arctic. All these states have formed The Arctic Council to ensure the sustainable development of the Arctic region (Brigham 2008; Pietri et al. 2008). In addition, there are many international governmental organizations and major powers from outside the area that have strategic interests in the Arctic region, mainly in the energy and mineral deposits that are becoming increasingly accessible as the ice cover retreats in the summer months (Brigham 2008; Hong 2012). Indeed, in May 2013, five non-Arctic countries, including China, Japan, and South Korea, were appointed as official observers of the Arctic area (Sakhuja 2013; Bennett 2014; Huang, Lasserre, and Alexeeva 2014). In addition, given the continuously increasing interest in developing shipping activity in and through the Arctic region, environmental concerns have come to the fore and there is now considerable importance attached to the need for regulating merchant ship operations in Polar waters (Brigham 2008; Jensen 2008; Pietri et al. 2008; Corbett et al. 2010; Molenaar 2014), not least from the Arctic states themselves (Wegge 2015).

2.2.4. Safety factors

These relate either to the potential risk of ships being involved in a marine accident or other unpredictable incident and/or for the crew to suffer injury and/or other adverse health consequences as a result of transiting Arctic waters. Safety factors can be divided into weather and geographic complexity, risk of crew health and safety, search and rescue, and environmental concerns:

• Poor visibility, fog and drifting ice are likely to increase the risk of accidents, especially since navigational data and information remain insufficient within the Arctic region (Lasserre and Pelletier 2011). The complexity of the geographical situation, together with impoverished charting, are also significant causes of concern for Arctic shipping routes (Meng, Zhang, and Xu 2017). However, Mussells, Dawson, and Howell (2017) point out that understanding and predicting the temporal and spatial distribution of ice is important for ship captains in the Arctic region, since their observations will be useful for the future development of viable seasonal shipping corridors and other safety-related policy decisions.

• Icing, darkness and low temperatures increase the risk of health and safety issues arising (Lasserre 2014). Floating ice in summer is still a threat to ships and the prevalence of ice in winter will obstruct the passage of most ships to some extent or other (Meng, Zhang, and Xu 2017).

• Search and rescue is difficult in polar areas due to the scarcity of accurate data (e.g. charts) (Ho 2010; Schøyen and Bråthen 2011; Lasserre 2014). Pruyn (2016) suggests that there is insufficient ice-breaking capacity within the NSR region to facilitate a swift rescue operation. Therefore, as Lammer (2010) points out, search and rescue coverage and capability within the region are rather limited (Lammer 2010).

• There are potential threats to the environment if ships are grounded or holed along the routes, especially in the case of oil tankers (Lasserre and Pelletier 2011; Hong 2012). The existence of these potential threats and impacts will necessitate a wide-ranging ecological risk assessment (Afenyo et al. 2016).

Using an approach based on Bayesian Networks, Afenyo et al. (2017) found that the most significant causal factors in Arctic shipping accidents were electronic failure of navigational equipment, mechanical failure of equipment (Navigational), failure of propulsion, human error (miscommunication), mechanical failure (communication equipment) and software malfunction.

In relation to avoiding the risk of damage to the ship, the use of ice-breakers as a support service and the presence of search and rescue services have been found to be critically important (Ho 2010; Farré et al. 2014). In addition, in order to avoid structural damage to the hull and propulsion system, the use of high ice-class ships (above the classic Baltic 1a or 1 AS) is recommended, especially during winter navigation (Bourbonnais and Lasserre 2015). Kum and Sahin (2015) applied root cause analysis to investigate Arctic marine accidents from 1993 to 2001. The results found that accidents to personnel are the most observed incident and that the negligence or carelessness of the injured person is the major root cause of such accidents (Lalone 2013).

3. Methodology

3.1. Justification of the Fuzzy Analytic Hierarchy Process (FAHP)

As revealed in the previous section, the decision to deploy a vessel onto an Arctic shipping route involves the consideration and balancing of numerous influential criteria. Within the context of the international shipping industry, this decision will most likely be taken at a corporate or strategic level, and most probably as a group decision.

The Fuzzy Analytic Hierarchy Process (FAHP) is a structured technique for organizing and analysing multi-criteria decision-making (MCDM) problems, where complex decisions need to be made in a context where there exist multiple objectives and multiple criteria which have an impact upon the decision(s). In applying this technique, fuzzy linguistic variables and associated fuzzy triangular numbers can be used for making comparisons among the influential attributes and, hence, provide solutions to vague and uncertain problems in decision-making (Zadeh 1965). As such, trade-offs and compromises need to be made, not only between the different objectives to be achieved (Saaty and Vargas 2012) but also, ultimately, between the various individuals that contribute to the joint group decision. It is with respect to this latter characteristic that FAHP exerts its most significant advantage over other multi-criteria decision-making techniques (Kabir and Hasin 2011).

FAHP and its related adaptations have been widely applied to the maritime industries. For example, based on FAHP, Celik, Deha Er, and Ozoke (2009) studied shipping registry selection in the Turkish maritime industry. Duru, Bulut, and Yoshida (2012) applied regime switching FAHP, utilizing direct numerical inputs to priority fuzzy sets, in order to solve the shipping investment strategy selection problem. Also, Chiu, Lin, and Ting (2014) used FAHP to evaluate green port factors and performance. Ding et al. (2014) adopted FAHP to evaluate key factors influencing new cross-strait shuttle shipping routes, while Sahin and Yip (2017) used an improved Gaussian FAHP model to analyse shipping technology selection.

Thus, FAHP has been selected as the most appropriate technique for the analysis of the Arctic shipping decision contained herein. Ultimately, this choice is justified because by combining fuzzy theory with the AHP method (Saaty 1980), fuzzy comparison matrices can be developed and dealt with. Such an approach improves upon the inherent weakness of AHP in that it assigns and utilizes exact numerical values in the judgement of comparisons, which proves to be rather ineffective when applied to ambiguous problems (Chang 1996).

3.2. The AHP model

The constituent steps in applying AHP are as follows:

3.2.1. Step 1: constructing the hierarchical model

Drawing upon both the literature review contained in the previous section, as well as a number of interviews with industry experts, a three-level hierarchical structure is developed. As illustrated in Figure 3, the AHP model encompasses the four criteria addressed in the previous section (i.e. economic, technical, political and safety factors). In addition, as Figure 3 also shows, a further sixteen sub-criteria are elaborated.

Figure 3. The structure of the AHP model.

The descriptions of the various sub-criteria and the sources which justify their potential influence (and inclusion in the model therefore) are shown for each of the key criteria in Tables 1–4. As can be seen in Figure 3, the criterion ‘Economic Factors’ includes five sub-criteria, namely: fuel costs, crew costs, insurance, other transit costs and sailing time. The ‘Technical Factors’ criterion comprises four sub-criteria, namely: ice breaker, ship construction, navigation and communication. The ‘Political Factors’ criterion consists of three sub-criteria, namely: right of navigation, maritime conventions and stakeholders’ concerns. Finally, the ‘Safety Factors’ criterion includes four sub-criteria, namely: weather and geographic complexity, risk of crew health and safety, search and rescue, and environmental concerns. As pointed out earlier in the paper, the allocation of specific sub-criteria to each of the four key criteria is based on the outcomes of an extensive review of the literature. It should be clear, however, that the potential for dependence between the four key criteria is quite significant; for example, the need for ice-strengthened ships (a technical factor) will obviously have an economic impact in terms of increased costs of ship construction and fuel use (Solakivi, Kiiski, and Ojala 2017). This potential problem of dependence between key criteria is largely overcome, however, by further disaggregation in the ensuing analysis to focus on the sub-criteria.

Table 1. Key ‘Economic Factors’ influencing the Arctic shipping decision.

Sub-criteria	Description	Sources
Fuel costs (F11)	Fuel consumption cost (including bunker, diesel and lubricating oil).	Laulajainen (2009); Ho (2010); Brubaker and Ragner (2010); Xu et al. (2011); Schøyen and Bråthen (2011); Hong (2012); Furuichi and Ostuka (2013); Lasserre (2014); Bourbonnais and Lasserre (2015); Lindstad, Bright, and Strømman (2016)
Crew costs (F12)	Costs related to officers and other crew members.	Liu and Kronbak (2010); Furuichi and Ostuka (2013); Lasserre (2014)
Insurance (F13)	H&M (hull and machinery), P&I (protection and indemnity).	Lasserre and Pelletier (2011); Xu et al. (2011); Furuichi and Ostuka (2013); Lasserre (2014)
Other transit costs (F14)	Other costs when transiting the routes.	Johnston (2002); Xu et al. (2011); Furuichi and Ostuka (2013)
Sailing time (F15)	Ship sailing time.	Lasserre and Pelletier (2011); Xu et al. (2011)

Table 2. Key ‘Technical Factors’ influencing the Arctic shipping decision.

Sub-criteria	Description	Sources
Ice breaker (F21)	Use of ice breakers to guide the vessels through shipping routes.	Xu et al. (2011)
Ship construction (F22)	Using an ice-class ship with thicker hull and more structural support.	Johnston (2002); Laulajainen (2009); Lasserre and Pelletier (2011); Hong (2012)
Navigation (F23)	Use of GPS//GLONASS (GLObal NAvigation Satellite System)/GMDSS(Global Maritime Distress and Safety System) to provide route recommendations.	Brubaker and Ragner (2010); Liu and Kronbak (2010); Schøyen and Bråthen 2011; Farré et al. (2014)
Communication (F24)	Use of communication technology to maintain clear message transmission (including voice and data transfer).	Ho (2010); Liu and Kronbak (2010); Farré et al. (2014)

Table 3. Key ‘Political Factors’ influencing the Arctic shipping decision.

Sub-criteria	Description	Sources
Coastal countries’ attitudes (F31)	Arctic shipping routes might be affected by coastal countries such as Russia and Canada.	Johnston (2002); Laulajainen (2009); Brubaker and Ragner (2010); Schøyen and Bråthen (2011); Hong (2012); Weber (2012)
Maritime conventions (F32)	Shipping right-of-way rules as established by the International Maritime Organization (IMO).	Pelletier and Lasserre (2012)
Stakeholder concerns (F33)	Stakeholders include Arctic and non-Arctic countries, shipping companies, port operators, governmental authorities, ship manufacturers, insurance companies, ship owners, shippers, legislators, logistics/warehouse service providers, environmental groups etc.	Johnston (2002); Ho (2010); Huang, Lasserre, and Alexeeva (2014)

Table 4. Key ‘Safety Factors’ influencing the Arctic shipping decision.

Sub-criteria	Description	Sources
Weather and geographic complexity (F41)	Harsh weather conditions (e.g. extreme coldness, fog, storms, total darkness) and drifting ice.	Lasserre and Pelletier (2011)
Risk of crew health and safety (F42)	Crews might fall ill in Arctic bad weather or suffer an accident.	Lasserre (2014)
Search and rescue (F43)	Search and rescue support services in case of emergency.	Ho 2010; Schøyen and Bråthen (2011); Lasserre (2014)
Environmental concerns (F44)	Environmental pollution from spill of oil, chemicals, ballast water etc.	Lasserre and Pelletier (2011); Hong (2012)

3.2.2. Step 2: constructing the pairwise comparison matrices

A survey questionnaire with a nine-point rating scale is used for study participants to express the relative importance they attach to paired evaluation factors. These responses on pairwise comparison scores were collected and collated to form pairwise comparison matrices.

3.2.3. Step 3: evaluating the weights of different hierarchies

The main elements of this step can be described as follows:

1. A reciprocal value is assigned to the inverse comparison:

(1)

where a_ij denotes the importance of the i^th (j^th) element.

• (2) Pairwise comparison in AHP is made within the framework of a matrix, and a local priority vector can be derived as an estimate of the relative importance associated with the elements (or components) being compared. This is achieved by solving the following equation:

(2)

…where A is the matrix of pairwise comparison, w is the eigenvector and is the largest eigenvalue of A (Saaty 1980). Therefore, A is consistent if and only if

3.3. The FAHP model

At this stage of the modelling process, the results of the AHP are combined with fuzzy numbers to form the FAHP results. The process can be described as follows:

3.3.1. Step 1: calculating fuzzy numbers

Zadeh (1965) defines a fuzzy set as a class of objects with a continuum of grades of membership ranging between zero and one. A triangular fuzzy number (TFN) is represented with three points as follows. A TFN is denoted simply as (l, m, u). The parameters l, m, and u denote the smallest possible value, the most promising value and the largest possible value, respectively. A triangular membership function is defined as Equation (3):

(3)

3.3.2. Step 2: building the fuzzy positive reciprocal matrix

The fuzzy positive reciprocal matrix may be expressed in the form:

(4)

…where the n elements of this matrix, , represent the pairwise comparison between criterion i and j using fuzzy numbers and where:

(5)

(6)

(7)

3.3.3. Step 3: calculating the fuzzy weights

The fuzzy weights of each criterion in the fuzzy positive reciprocal comparison matrix are obtained by utilizing the geometric mean method (Buckley 1985), where the geometric mean of the fuzzy comparison value of criterion i to each other criterion is derived as follows:

(8)

The fuzzy weight of the i^th criterion, as represented by a TFN, is then computed as follows:

(9)

3.3.4. Step 4: defuzzification

Because the criteria weights are still in the form of fuzzy triangular values, there is a need to translate them into non-fuzzy values through a process of defuzzification. According to past studies (e.g. Van Leekwijck and Kerre 1999; Hsieh, Lu, and Tzeng 2004; Chou and Chang 2008; Yeh and Xing 2016), defuzzification can locate the best non-fuzzy performance (BNP) value based on the centre of the area or centroid. The BNP value of a fuzzy number is obtained as follows:

(10)

3.3.5. Step 5: normalization and synthetic analysis

Although the previous step yields a non-fuzzy number, in order to compare the relative importance attached to each of the evaluation criteria and sub-criteria analysed within this study, there is now a need to normalize these BNP values for a given criterion (or sub-criterion) as follows:

(11)

3.4. Data collection

In order to apply the FAHP model to the specific context under study, data are required not only on the factors influencing the Arctic shipping decision, but also on the perceptions amongst the specific decision-making community concerning the relative importance of these factors and how they might be traded-off against each other. To this end, a questionnaire survey was developed to elicit appropriate responses from senior executives of shipping companies based in Taiwan. This geographical context constitutes one of the major centres for shipping activity within Asia and, as such, many international shipping operators are located within Taiwan, thus yielding the potential for quite a large, but easily accessible, sample. In addition, Taiwan is pivotally located for connecting into both the NSR and the NWP and, therefore, could become one of the primary origins or destinations for future Arctic freight movements. The sampling frame was all shipping companies based in Taiwan whose operations were such that they might have the possibility of utilizing an Arctic shipping route. The target companies identified are listed in Table 5. Given the corporate or group nature of strategic decision-making within the identified target companies, the next stage was to identify specific individual senior executives within each of the companies that would be a potentially relevant and useful respondent to the survey.

Table 5. Taiwanese shipping companies identified within the sampling frame.

Evergreen Marine Crop. Ltd. Yang Ming Line Wan Hai Lines T.S. Line Hanjin Pacific Corp. Chinese Maritime Transport Ltd. NYK Line COSCO Shipping Cheng Lie Navigation (CNC) Line Orient Overseas Container Line Taiwan Navigation Corp. Ltd.

Once the survey questionnaire was designed, it was piloted with five local experts in the field in order not only to validate its content, but also to further refine the wording of the questionnaire so as to improve its readability. Ultimately, the questionnaire survey was administered to 28 senior shipping experts who understand Arctic shipping issues and work in shipping companies based in Taiwan. The questionnaire was sent to each of these executives by mail with a postage-paid return envelope. Before sending the questionnaire, potential respondents were contacted by telephone to obtain their agreement to participate in the study.

The majority of the questionnaire content revolved around the use of a nine-point rating scale to measure the respondents’ perceptions as to which of four factors and sixteen attributes they considered to be relatively ‘important’ and relatively ‘unimportant’. In addition, objective data on the respondents’ basic characteristics (such as job title, age, educational level, length of time with company and position within company) was also collected.

4. Results

By the cut-off date for the return of completed questionnaires, 25 usable questionnaires had been received. For each response, the consistency index (CI) was calculated and tested to confirm the consistency of its pairwise comparison matrix. The results indicated that two responses with a CI of less than 0.1 were highly inconsistent (Saaty 1980) and were consequently discarded. Therefore, the overall response rate was 82.1% (=23/28). A profile of the characteristics of the 23 respondents is shown in Table 6. It can be seen that most of the respondents are senior executives with at least 10 years working experience in the shipping industry, thus indicating a high level of reliability of the survey findings.

Table 6. Profile of the survey respondents.

Characteristic	Range	Frequency	Percentage (%)
Job title	President/Director	3	13.04
	Senior deputy director	5	21.74
	Division director	5	21.74
	Supervisor	4	17.39
	Senior engineer	6	26.09
Age (years)	Under 40	4	17.39
	41~50	4	17.39
	51~60	13	56.52
	Above 60	2	8.70
Educational Level	Bachelor	17	73.91
	Master	6	26.09
Seniority	10~15	4	17.39
	16~20	8	34.78
	21~25	5	21.74
	Above 26	6	26.09

Based on the AHP results, the local weights of each construct and influencing factor are shown in Table 7. All the consistency ratio (CR) values are less than 0.1, and thus meet the requirement of the consistency test. The results indicate that economic (0.3045) is the most important factor influencing the choice of Arctic shipping, followed by safety (0.2894), technical (0.2174), and political (0.1887). With regard to attributes, fuel costs (0.2374), communication (0.2719), right of navigation (0.3608), and risk of crew health and safety (0.2798) were perceived to be the most important sub-criteria with respect to each of the economic, technical, political and safety factors, respectively.

Table 7. Results of the FAHP analysis.

Factor	Local weights	Consistency ratio (CR)	Attributes	Local weights	Global weights	Rank
Economic	0.3045	0.01724	Fuel costs	0.2374 0.1897 0.2007 0.1991 0.1730	0.0723 0.0578 0.0611 0.0606 0.0527	3 12 7 8 14
			Crew costs Insurance Other transit costs Sailing Time
Technical	0.2174	0.06897	Ice breaker	0.2583 0.2298 0.2400 0.2719	0.0562 0.0500 0.0522 0.0591	13 16 15 11
			Ship construction Navigation Communication
Political Climate	0.1887	0.00172	Right of navigation	0.3608	0.0681	5
			Maritime conventions	0.3190	0.0602	10
			Stakeholder concerns	0.3201	0.0604	9
Safety	0.2894	0.08621	Weather and geographic complexity	0.2522	0.0730	2
			Risk of crew health and safety	0.2798	0.0810	1
			Search and rescue	0.2309	0.0668	6
			Environmental concerns	0.2372	0.0686	4

The analysis progresses by determining the global weights for the second level by multiplying the local weights and the corresponding criteria in the level above, and adding each of these to each element in the level according to the criteria considered. The results reveal that the top three most important sub-criteria influencing the choice of Arctic shipping are risk of crew health and safety (0.0810), weather and geographic complexity (0.0730) and fuel costs (0.0723), respectively.

5. Discussion and conclusions

The most obvious and confirmatory conclusion to draw from the results of this analysis is that ‘Economic Factors’ have again been found to be significant in the choice of whether or not to utilize the possibility of Arctic shipping. The distance and time savings of Arctic shipping compared to conventional routes are obvious and well understood (Beveridge et al. 2016), yielding as they do quite significant fuel savings. As one might expect, fuel costs are indeed found to be the most important influence on the decision within the category of ‘Economic Factors’. However, the influence of fuel costs is only ranked 3 overall with respect to the sub-criteria. A possible explanation for this lies with the fact that numerous studies have found that the scale of the fuel savings associated with Arctic shipping routes is not commensurate with that of the distance and time savings and that increases in other direct costs go some way to counteract this most obvious potential benefit of selecting an Arctic route (e.g. Verny and Grigentin 2009; Somanathan, Flynn, and Szymanski 2009; Liu and Kronbak 2010; Schøyen and Bråthen 2011; Kiiski 2017). This is also clearly indicated by the relatively high ranking attached to the ‘Economic Factors’ sub-criterion of ‘insurance’ (ranked 7 overall), because of the fact that premiums for insurance cover will certainly increase with greater exposure to the higher risks associated with Arctic shipping (UK P&I Club 2016) and the fact that the cost of H&M insurance will certainly increase relative to the cost of P&I insurance (Meng, Zhang, and Xu 2017). There is evidence to suggest, however, that the scale of additional insurance costs may pose more of a problem in perception rather than actuality, and that this might provide a perceptual deterrent to the expansion of Arctic shipping movements (Sarrabezoles, Lasserre, and Hagouagn’rin 2016).

Other transit costs are ranked as the third-most important influence within the category of ‘Economic Factors’ (ranked 8 overall). Of course, the actual value payable depends on types of ship and cargo (e.g. liner or tramp service), ship class and size, ice conditions and sailing routes, etc. Also, in evaluating these ‘other transit costs’ of Arctic shipping against those of traditional routes, any changes to the pricing policies of the Suez and Panama canals will exert a significant influence.

The other highly ranked sub-criteria resulting from the analysis provide a significant insight into the less obvious influences on the Arctic shipping decision. The top two most highly ranked criteria emphasize the paramount importance of ‘Safety Factors’ (with ‘risk of crew health and safety’ ranked as 1 overall and ‘weather and geographical complexity’ ranked as 2 overall. All of these highly ranked sub-criteria relate in one way or another to the greater risks associated with opting for an Arctic shipping route and implies quite a high degree of risk aversion amongst survey respondents in their assessments of the potential factors influencing the Arctic shipping decision. This result could be due to the prevalence of container shipping operators within the sample surveyed, which is a representative characteristic of the sample population in Taiwan. As evidence exists that bulk shipping operators are less risk averse than their container shipping counterparts (Lorange and Norman 1973; Cullinane 1991; Kavussanos and Alizadeh 2002;   Adland and Cullinane 2005: Tezuka, Ishii, and Ishizaka 2012), perhaps these risks would be less influential where the sample is not dominated by the latter group of decision makers; an aspect which is worthy of consideration in future research based on a different sample.

The results of this study suggest that the risk factors, mainly in the ’Safety’ and ’Political’ domains, exert a far greater negative influence on the propensity to deploy vessels on Arctic routes than the supposed positive influence of ‘Economic’ factors, particularly as manifest in reduced fuel use and reduced transit time. This implies a certain lack of appetite amongst shipping operators to opt for Arctic shipping as a viable alternative to conventional routes. These findings represent a novel and significant contribution of this work. Through the collection of detailed primary data and the original and rigorous application of an appropriate methodology at both aggregate and disaggregate level, conclusions have been drawn which contradict, or at least qualify, the findings of previous work with respect to the overriding influence of economic (cost) considerations. Ultimately, the risks involved are found to exert much more of a deterrent effect than the attraction of reduced costs. This conclusion does go some way to confirming that of Lasserre and Pelletier (2011) who surveyed the intentions of 98 shipping operators and concluded that there was significant reticence amongst them towards the Arctic shipping option.

Lasserre and Pelletier (2011) further identified that while bulk shipping operators were highly cautious in their intentions to use Arctic shipping routes, the container sector were completely disinterested. Given the requirement for standardized schedules in container shipping, as well as the need for a highly formalized and integrated network structure (Wang and Cullinane 2008; Cullinane and Wang 2009), this does not come as any great surprise. Indeed, it appears to provide strong support for the assertion of the Arctic Council (2009) that, rather than transit traffic linking Europe and the Americas to Asia, the major source of increasing Arctic traffic will relate to bulk shipping movements associated with serving Arctic ports and the exploitation of natural resources as, into the future, they become increasingly more accessible with further shrinkage in the polar ice cap (Arctic Council 2009).

There is no doubt that the continuing shrinkage and thinning of the polar ice cap will certainly improve the economics of the Arctic routes, as the risks and difficulties associated with Arctic transits are reduced. Bathymetric considerations will still provide a key constraint on the growth of Arctic shipping, since the size of ship which may be deployed is somewhat restricted (Meng, Zhang, and Xu 2017). However, the times of the year when transits are practically feasible will expand and there will be greater potential than previously for more direct routes in deeper waters using bigger ships.

This likely future scenario will inevitably further supplement the expansion of Arctic shipping movements which has already been seen over the past decade. It is unlikely, however, that the economics of Arctic shipping will ever prompt the re-routing of liner shipping services (Kiiski 2017) and that future growth in Arctic shipping will remain the preserve of bulk shipping. Future demand will not be primarily motivated, therefore, by the desire to transit Arctic waters to establish more efficient bulk shipping links between Asia and the West, but rather for bulk shipping to service the ports, mines and wells that will increasingly be located within the Arctic itself. Another major contribution of the analysis undertaken herein is to specifically identify the risks (or barriers) that need to be reduced (or overcome) in order to facilitate the future expansion of Arctic shipping to take advantage of the ever-improving economics associated with it. In doing so, however, it is important to recognize that external influences beyond those explicitly considered herein can have an important influence over the future development of Arctic shipping.

Acknowledgements

The authors are grateful to three anonymous reviewers who provided feedback on an earlier version of this paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. The Transpolar Sea Route is not discussed in this paper.