Investigation of empty container shortage based on SWARA-ARAS methods in the COVID-19 era

In container transportation operations, uncertainty in demand can be experienced periodically. Maritime companies have developed several of operational and strategic practices to deal with these situations when they occur, including omitting some ports in the itineraries or the blank sailing of all ports in the itineraries and assigning ship to schedules on different routes [18]. These practices used as solutions by maritime companies can, however, result in empty containers failing to be positioned at the necessary points [19] and thus lead to redundant storage of containers in regions where importation is high, as well as to container shortages in regions where exportation is high [20]. This issue was addressed in a study conducted by Ko [21], where it was highlighted that in countries with high global trade volumes, there is considerable demand for empty containers, and when the demand cannot be met, the available-to-promise cargoes are kept in warehouses. Empty containers that are rendered out of use by keeping them in warehouses or ports cause both cost and time losses for maritime companies, as the storage cost of these containers is almost equivalent to that of full containers. Moreover, when including the use of operational equipment and personnel costs to ship these containers, the total cost can run much higher even [22].

Seven of the world’s 10 largest container ports are located in China, and more than 50% of the global container shipping volume is handled at Chinese ports, the main reason being that China is not only the world’s largest exporter but also the world’s second-largest importing country. This global structure governing exports and imports necessitates the formation of a China-oriented container positioning cycle in container transportation. After a full container is discharged according to this positioning cycle, the container needs to be quickly positioned in China, regardless of whether it is empty or full and then shipped to different regions once it is full. This cycle plan increases the number of containers exported from China and leads to an insufficient number of imported containers. In addition, trade volumes of containerized cargoes increase rapidly with the increase in China-based trade activity that typically occurs before the New Year [14]. The empty container shortage recently experienced can be attributed to all the commercial activities carried out in this process [22]. The emergence of COVID-19 in China, which is recognized as the export center of the world, during the period of the highest export volumes resulted in a rapid decrease of commercial activity in the country and a much more intensive container shortage problem in the world than ever seen in the past [23]. This problem was analyzed by Xu [24] with a regression model using data obtained from 14 different Chinese ports. The results of their study revealed that the port operation processes for both export and import cargoes in China were adversely affected by COVID-19, being responsible for a 20–50% decrease in cargo volumes at Chinese ports [25].

After the declaration of COVID-19 as a pandemic, global-scale quarantine and restrictive practices were applied all around the world. These practices have also led to disruptions in global-scale supply chain networks [26], as the national and international quarantine and restriction policies imposed in the first half of 2020 have resulted in many countries experiencing foreign exchange and labor shortages and sharp decreases in consumer demand [27]. Within this period, 255 million full-time employees working in global-scale operations have been laid off [28]. The most intensive impact of the pandemic on global trade volumes was observed in the second quarter of 2020, which witnessed a 21% decrease in value-based global export and import volumes [29]. More than 80% of international trade is carried out by maritime transport [2], a global trade figure that suggests a problem in maritime transport will also have an impact on global trade. Indeed, in a study by Verschuur et al. [30] conducted using ship monitoring data, it was concluded that volume losses of between 206 and 286 million tons in global maritime trade occurred in the first eight months of 2020. When shipping companies do not get sufficient bookings from ports in their regions of operation due to decreased trade volumes they reduce their port calls or declare blank sailing until the demand in the region reaches desired levels. In the second quarter of 2020, there was a 17% decrease in the number of ports of call [31], and in the first half of 2020, the number of ships actively operating decreased by 69% compared to that of 2017 [32]. In a study conducted by Narasimha et al. [33] that compared quantitative performance data from the COVID-19 and pre-COVID-19 period, it was determined that there was a decrease in cargo volumes and the number of ports of call. In the same study, the effects of COVID-19 on maritime operations were examined using the data obtained from 87 maritime industry experts, where the results showed that the three most common problems were associated with the workforce, the decrease in load volumes, and operational delays, in descending order of magnitude. During this period, it was announced that among three key shipping alliances, there were 126 blank sailings on the Trans-Pacific trade route and 94 on the Asia-Europe route [34]. In the period leading to October 2020, there were 515 blank sailings [35]. In the Suez Canal, one of the most important transit routes on the Far Eastern Europe route, the number of transits in May decreased by 32% to reach an all-time low [36]. In 2021, shipping companies announced “blank sailing” for 919 container ships on Transpacific and Asia-Europe routes and decided to temporarily suspend the port calls of the ships [37]. Container transportation is also defined as regular line transportation because itineraries, price tariffs, and ports of call are determined before the voyage [38]. Most maritime companies publish their schedules, including the ports of call on the service route, months in advance [19], which means that when the itinerary of container transportation is not executed as planned, this decreases the reliability level of maritime companies [39]. Although time losses that affect fixed schedules are experienced due to weather conditions, port planning, and problems in operational processes, these losses are tolerated by customers. However, as of 2020, sharp declines in the reliability of itineraries and problems caused by container shortages have decreased the reliability levels for container line operators considerably. In a report investigating the reliability of itineraries using data obtained from 34 trade routes and more than 60 maritime routes, it was determined that the reliability decreased to 34.9% in the first quarter of 2021, the lowest level since 2011 [40]. During this period, the precautionary strategies applied by the maritime companies were not accepted by the BCO, and complaints were made to many different commissions, such as the FMC (Federal Maritime Commission). A majority of the complaints involved the decreased number of sailings (navigations), blank sailing advertisements, low level of schedule reliability, container rollover, additional costs, and container shortage [35].

The rapid decline in global trade volumes caused by COVID-19 also triggered significant time losses in port operations [33]. This is because the variation in international trade volumes is one of the most important variables that directly affect the efficiency of container ports [24]. Significant disruptions particularly occurred at the entrances and exits of highway and water transportation connections to ports after COVID-19 was declared a pandemic [41]. In addition, the restrictions imposed during this period have brought basic logistics activities, such as inland transport, customs, and warehousing, to a standstill and have caused empty containers to remain at the ports in Europe and America, leading to container shortage problems in other regions [42]. In the first half of 2020, the quarantines imposed at the ports and the limited number of personnel available to provide services increased the costs of container transportation [26]. The impact of all these negative developments can be seen in the China Containerized Freight Index as a 50% increase [43]. In the first three quarters of 2021, 40 ft container freights increased by 477% on the China-US route and by 243% on the China–Europe route [44].

COVID-19 has led to disruptions in global supply chain connections, causing major problems in international trade [45]. In the study by Cengiz & Turan [46] carried out with the participation of experts from 21 countries to examine the effect of COVID-19 on maritime transport, it was determined that BCO companies faced major challenges in their import processes. Particularly in the case of the uncertainty of delivery times, companies are unable to make plans about raw materials and production and have to deal with unanticipated financial losses in supply chain processes, such as packaging, storage and distribution. In one study, it was reported that 94% of the companies on the Fortune 1000 have experienced major problems in their supply chain connections due to issues of uncertainty in supply chain operations caused by the pandemic [47]. One of the main methods applied by maritime companies to address the delays in both inland transportation and in-port transfers of export cargoes has been to reduce the "Free Time" of containers. Some BCOs choose to leave their cargoes at the terminals due to short free times and higher expiration fees, such as demurrage and detention costs [48].

To address the problems in maritime transportation, many maritime lines have chosen to increase the end-to-end logistics service networks that they offer to foreign trade companies through horizontal and vertical integration expansion [41]. Because the container transport chain comprises several different business processes; therefore, to move a container from one point to another, many independent parties must continuously communicate with each other. For example, avocado shipment in containers between Mombasa and Rotterdam encompasses 30 independent parties and 100 employees in the process and conducted 200 information transfers [49]. Shipping lines provide logistical services to prevent communication gaps in the chain and direct communication with the shipper during all processes, including the shipment and delivery of the cargo to the customer (E2E), rather than just providing container transportation services. For example, A.P. Moller-Maersk, which has the world’s largest container-ship fleet, provides E2E logistical services, including major services, such as inland transport, warehousing, customs, and distribution, in addition to container-transportation service [50].

Due to the low level of demand for container transportation, containers that have already been booked are kept in warehouses for some period of time and bookings for the subsequent days increase the demand. Thus, ships cannot be available and often containers are rolled over. Many strategies have been developed by shipping lines to resolve this issue. The practice of shipping guarantee booking developed by Hapag-Lloyd is one of these strategies. Thanks to this practice, the company helps its customers who struggle against uncertainties in the supply chain by preventing container rollovers during the uncertainty period caused by COVID-19, especially during the high-volume raw material import period [51].

When empty containers cannot be supplied, purchasing or renting containers can be an important resolution to the container-shortage issue. In a recent study, Ko [21] has concluded that purchasing containers is much more cost-effective than renting. In addition, there are advantages to purchasing containers for the shipper, such as the elimination of demurrage and detention costs, which would normally be paid to the container shipping company because of excess free time.

Another method developed to overcome the insufficiency of empty containers is the inducement call. The demand uncertainty because of communication gaps is among the main reasons for the container shortage. For example, Maersk announced periodical blank sailing in services where the demands decreased because of COVID-19 and then decided to make an inducement call to the ports where there was sufficient demand for the ships operating within the canceled route [52]. Thanks to this, BCOs could continue their activities within regions of high demand. In this process, many methods have been tried in order to follow customer demands and eliminate communication deficiencies. Logistics information systems-based technologies were insufficient to meet the needs in this process. Information and communication technologies, which have provided digital visibility, must be redesigned with the technology of the autonomous age to provide resistance to the container-transportation chain against the difficulties caused by COVID-19. Information communication technologies redesigned with Industry 4.0, a new phase in the Industrial Revolution that provides digital visibility to managers of companies in the container-transportation chain and enables large data to be collected, transmitted, and processed at one point, presents real-time visibility and decision-making support by optimizing the entire transportation chain within the virtual environment [53]. Accordingly, shipping companies benefit from the block chain infrastructure, which is an important achievement of Industry 4.0, to achieve a continuous connection within the container-transportation chain during COVID-19 [54]. The Mediterranean Shipping Company (MSC) has launched the “electronic Bill of Lading” (eB/L) application in Wave BL with block chain technology to enable continuous communication among the parties in the chain [55]. Another important example within the scope of information and communication technologies is the “spot booking application”, which was developed before COVID-19 and launched by the A.P. Moller-Maersk Group, which offers a web-based solution to BCOs for global uncertainties during the pandemic period. This application is an information communication technology that allows BCOs to create a reservation record for a ship without contacting a company representative at the reservation stage [56].

SWARA is a decision-making method developed by Keršuliene et al. [58]. One of the advantages of this method, as compared to both simple and other multi-criteria decision methods, is that it provides the opportunity to obtain the same results with a limited number of mathematical operations [59]. For example, Erdoğan et al. [60] demonstrated that the same results for determining criterion weights as those achieved using the ANP (Analytic Network Process) can be achieved with the SWARA method using fewer mathematical operations. More specifically, it was shown in one case that while 67 binary comparisons were made in ANP, 9 comparisons were made in SWARA, with similar results. In a similar study carried out by Stanujkic et al. [61], SWARA and AHP (analytic hierarchy process) methods were compared, and it was determined that it was easier to calculate mathematical operations by making less comparisons with the SWARA method, as compared to the AHP method. The SWARA method is appropriate for solving the decision problem where the importance level of the criteria is known [62]. In addition, this method allows the experts who created the data set to freely express their feelings, thoughts, and opinions without being dependent on external factors. From the data obtained from sectoral experts, it is possible to solve the incompatibility problem and create rational decision models by examining the criterion weights [63]. Therefore, the decision hierarchy formed as a result of the data analysis is seen as the numerical equivalent of the experts’ experiences [58]. On the other hand, the main disadvantage of the SWARA method over other MCDM methods is that the calculation procedure does not include a procedure for determining the consistency of pairwise comparison [61].

Examination of the literature shows that the SWARA method has been used to come up with solutions to many decision-making problems. For example, in a study by Prajapati et al. [64],​the relative impact of barriers to reverse logistics implementation was evaluated using SWARA, while in a study by Agarwal et al. [65], the ranking of criteria related to the creation of barriers to human supply chain management was evaluated by using fuzzy SWARA to improve it. The SWARA method has also been applied to rank the problems caused by the lack of sustainable suppliers in the electronics industry [66]. Moreover, Yücenur and Ipekçi [67], in the model they proposed, ranked the criteria for a site selection problem by using SWARA for the first offshore current power generation facility planned to be established in Turkey, while Baç [68], who evaluated different smart card systems, used the SWARA method to determine the best alternative and the criteria weights in the decision model.

With the SWARA method, the weights are determined using the following six steps:

To reduce the criterion weight determined by each decision maker to a single value, integration is performed by taking the arithmetic average of the calculated weight of each decision maker for the relevant criterion, after which the final criterion weight is obtained.

The additive ratio assessment (ARAS) method was developed in 2010 to provide a new approach for resolving multi criteria decision-making issues [69]. This method was developed to calculate the utility ratios of the alternatives and to rank the most appropriate alternatives using simple mathematical calculations [70]. The procedures used for the mathematical calculations of the ARAS method are simpler than prominent MCDM methods, such as TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution), VIKOR (VIse KriterijumsaOptimiz acija I Kompromisno Resenje), and PROMETHEE (Preference ranking organization method for enrichment evaluation) [71]. In MCDM methods, it is very important to use the utility function to evaluate the decision makers and to create the numerical equivalents of all combinations of alternatives. This is where the ARAS method has a major advantage, insofar as the alternatives in the decision hierarchy are ranked according to their utility function values [72] and compared with the utility function value of the optimal alternative determined by the decision maker. In other words, the utility function value in the ARAS method is equivalent to the relative effect of the values and weights of the alternatives. This method allows complex problems to be simplified and the most appropriate alternatives to be ranked with their utility ratios, without using different calculation tools [73]. To simplify, by determining the utility function value, the most appropriate alternatives can be ranked [69]. It is possible to determine this value by proving other independent conditions after finding the multi-attribute utility function and then creating multi-attribute utility functions [74]. This provides compatibility in many different sectors due to the fact that it yields accurate results in terms of determining easy, understandable, and most appropriate alternatives [58]. The ARAS method has been used in more than 95 scientific articles between 2010 and 2020 due to its advantages [73]. These articles used the ARAS method to determine everything from underground site selection for hydrogen storage [75] and calculation of sustainability indicators for renewable energy systems [58] to concept selection for load distribution [76], the most appropriate personnel selection [77] and accountant selection [78].

Resolving the decision-making issues using the ARAS method entails the following four steps:

In the X decision matrix as shown in Eq. (4), m is as the number of alternatives and n is the number of criteria; x_ij is the performance value of ith alternative in the jth criterion, while x_0j is the optimal value of the jth criterion using Eq. (4).

For the decision-making issue, when there is no information about the optimal value of the relevant criterion, the optimal value is calculated by using Eq. (5) when the criterion shows the utility attribute or using Eq. (6) when it shows the cost attribute.

When \({\bar{x}}_{ij}\) values show the cost attribute, the cost function is converted to the utility function using Eq. (8) and the normalized value is calculated using Eq. (9).

After calculating the normalized values of the criterion performance values, the normalized decision matrix \(\overline{X}\) is obtained using Eq. (10).

The members of \(\hat{X}\) weighted normalized decision matrix \(\hat{x}_{ij}\) are obtained as shown in Eq. (12).

Weighted normalized decision matrix \(\hat{X}\) is obtained in the matrix form given in Eq. (13).

The value > \({S}_{i}\) calculated for each alternative indicates the most effective alternative for the decision-making issue. Using Eq. (15), \({K}_{i}\) utility ratios are calculated by proportioning \({S}_{i}\) values of the alternatives to the \({S}_{0}\) optimal function value.

The effects of the alternatives on resolving the decision-making issue can be calculated using the \({K}_{i}\) utility ratios within the range of [0, 1]. The alternatives can be assessed by ranking the utility ratios. By using the \({K}_{i}\) ratios, the relative efficiency of the utility function values of the alternatives are calculated. Accordingly, the alternatives are then assessed by ordering the values.

In the first stage of the study, the literature was reviewed to determine the effects of the empty-container charge on container transportation and offer solutions, after which decision-making issues were identified using the opinions of 19 experts working in BCO and FFW companies that represent the demand for the container-transportation chain. In the second stage, criteria weights were calculated by evaluating the criteria determined by the experts. In the last stage, solutions were evaluated by experts using a range of 10–100 points according to the determined criteria, the optimum resolution proposal was calculated, and the study was concluded. The flowchart of the study is shown in Fig. 1.

The main criteria and subcriteria used as data collection protocols in the present study were constructed on the basis of studies reviewed in the literature, sectorial publications, and expert opinions. The criterion sampling technique was used to determine the group of experts to be involved in the study. Using this technique, the experts having sufficient experience and knowledge by which to evaluate the container-shortage issues were assessed using a set of questions (e.g., university graduation, sectorial experience, department, and position) and then included in the study. To evaluate the issues caused by the empty-container shortage on the demand (customer) side, the experts comprised 19 managers working in BCO and FFW companies. The sector representatives of the study group were interviewed through face-to-face and online applications.

To combine the sectorial experience with the theoretical knowledge to determine the expert group, a prerequisite was to have an associate degree/undergraduate education related to seaborne trade/transportation. Of all participants, 36.84% had completed their education in maritime business management, 31.58% in foreign trade, and 31.58% in international trade and logistics. In addition, 42.11% of the company representatives participating in the present study had 4–6 years of sectorial experience, 36.84% had 7–9 years, and 21.05% had 10–12 years (Table 1).

The precautionary strategies conducted to reduce the effects of COVID-19 have caused restrictive effects on global trade and major disruptions in labor productivity, production, and distribution processes and are considered in container transportation, the building block of global trade, as the container-shortage problem [79]. To reduce the negative effects of this problem on the container-transportation chain, many strategies have been developed during the COVID-19 pandemic. The issues within the container-transportation chain, the explanations of the criteria, and their abbreviations used in the calculation tables are presented in Table 2, while the suggestions to resolve these issues and their abbreviations are presented in the decision hierarchies in Table 3. In addition to the information obtained from the literature review, the main criteria and sub-criteria used in the decision hierarchy were determined, taking into account the suggestions of the experts working in the sector.

The criteria weights were calculated using the SWARA method based on the knowledge and experience of each expert in the Decision Committee. In order not to prolong the study, only the findings of the first expert are presented in Table 4 as an example.

The criteria weights were calculated in line with the opinions of experts working in BCO and FFW companies. The criterion weights of the decision committee were found by averaging the weights using the arithmetic mean operator in order for the weights of the calculated criteria to express the consensus opinion of the decision committee. The ranking of all criteria according to their weights and importance levels is given in Table 5.

The main criteria for the empty container shortage were ranked as follows based on their importance levels: CTCI (35.04%), SCU (24.98%), VLCT (22.11%), and BSAI (18.88%). The main criteria were then ranked according to their importance as follows: NI (13.22%) in CTCI, HAPL (6.46%) in SCU, ITDI (5.89%) in VLCT, and UIS (5.70%) in BSAI. When the general rankings were created according to the importance levels among the criteria, the first five subcriteria were as follows: NI (13.22%), DDCI (7.86%), DII (7.58%), HAPL (6.46), and ITCI (6.36%).

To determine the most appropriate method among the suggested alternatives, the evaluations between 10 and 100 points of the experts participating in the study were taken as a reference. A decision matrix was created using the geometric mean to express the common opinion. The results of the calculated group decision matrix are shown in Table 6, while the results of the group decision matrix weighted for the alternatives are shown in Table 7.

The criteria weights calculated using SWARA and the weighted group decision matrix are shown in Table 7.

The optimum function, utility ratios, and resolution rankings calculated for the demand sector are presented in Table 8. When examining the findings, SGB was determined to be the most beneficial alternative with a 95.08% utility ratio among all alternatives proposed for the empty container shortage. The utility ratios of other resolution alternatives suggested in the study were ICT (88.28%), SOC (87.60%), IC (84.88%), and E2EDS (62.93%).