Supply chain resilience in the Colombian defense sector before and during the COVID-19 pandemic: A comparative study

Unforeseen events can significantly affect organizations’ supply chains and disrupt their continuous flow. Therefore, organizations need to develop a response capability that allows them to minimize the negative effect of such events and quickly recover from them, also known as resilience. This research performs a comparative analysis of the influence that risk, vulnerability, and adaptability have on the resilience capability of supply chains in Colombian defense sector organizations before and during the coronavirus outbreak. Based on a literature review, a survey was designed and applied online to collect data from respondents related to the activities of the Colombian Air Force supply chain. For the first wave, data was collected between December 2019 and January 2020. Data for the second wave was collected in August 2020. Results suggest that identifying and managing risks positively impact reducing vulnerability and increasing adaptability. Moreover, by decreasing exposure and improving adaptability, the organization positively influences supply chain resilience capability. The results also indicate that the pandemic positively affected risk and vulnerability awareness. The identification of vulnerabilities had a positive impact on the resilience capacity during the Corona Virus outbreak. This research provides relevant information for the Colombian government on developing public policies and mechanisms of service and support for defense sector organizations to strengthen their resilience capability. Likewise, the study offers valuable information to those organizations interested in improving their resilience capability and that of the sector in which they are involved.


Introduction
This research began in mid-2019 by analyzing the supply chain in the defense sector in Colombia. Specifically, the vulnerability, the risk, and the adaptability influence the supply chain resilience capability. At the beginning of 2020, after analyzing the collected information, the World Health Organization declared that the coronavirus outbreak had become a pandemic on March 11 th . This event impelled a reassessment of the research path. As a result, it was decided Because of market globalization, supply chain and logistic processes are becoming more complex, especially for Colombian organizations, which must deal with limited infrastructure, political unrest, security issues, and rugged topography. Additionally, several organizations with world-class supply chain practices compete in the same market. Consequently, Colombian organizations need to learn about the commercial conditions, logistics, and supply chain operations in the global and local markets. This helps improve their competitive position and resilience in the disruption of the supply chain [17].
Independent of how well-managed the supply chain is (coordinated, planned, implemented, and controlled), unexpected events may affect the process's efficiency. The COVID-19 pandemic was an incredible event-considered the most severe global crisis since the second world war-and it is pretty sure that it will have a long-term impact on economies, supply chains, and trade relations. The different governmental containment measures have caused severe disruptions in the global supply chains [18], affecting the supply of components and raw materials manufactured in countries such as China, Vietnam, and India, that act as supply chain nodes because they hold massive and inexpensive materials and labor sources. This exposed vulnerabilities in numerous industries and sectors with no alternatives to switch rapidly to other sources [19]. This situation may encourage organizations to rethink their supply chain strategies by diversifying and having alternative suppliers [18].
Therefore, it should be an organizational priority to identify the factors that can affect its supply chain the most and try to prevent and handle potential disruptions. Although there are several studies in the literature analyzing factors that affect the supply chain, most of them have been carried out in developed countries, and few of them are focused on government institutions [16,[20][21][22][23][24][25][26]. Furthermore, there is no evidence of studies focused on Colombian defense organizations that consider the effect of risk, vulnerability, and adaptability on the supply chain's resilience, which are the factors examined in this study. Therefore, the question addressed in this research is: "What are the influence of risk, vulnerability, and adaptability on the supply chain resilience capability in organizations in the defense sector in Colombia before and during the coronavirus outbreak?" To answer the research question, based on a review of the literature, a reliable instrument was designed to collect data from the Colombian Air Force, which has several facilities, bases of operations, and a vast network of suppliers. This study based its inferences on analyzing the interactions of the selected factors in the chosen institution at two different times: just before the coronavirus outbreak and during the pandemic. The Structural Equation Modeling (SEM) technique was used to analyze the collected information.

Supply chain in the defense sector
The worldwide SCM's best practices in the private sector have been adopted as a benchmark to assess the efficiency of organizations in other sectors in managing their supply chains. Nevertheless, using SCM best practices in public organizations is still incipient since few articles in the literature deal with SCM in the public sector [27][28][29][30]. It may be naïve (or irresponsible) to make a list of industries and choose a supply chain 'leader' for all of them. Each industrial sector is unique and has its complexity when analyzing the supply chain [31]. Among those many sectors, defense is one of the most challenging.
For the defense sector, the mission of SCM must be to provide comprehensive logistical support for the different weapons and personnel. This support should be precise in its operation but minimize the cost of providing it [32]. Therefore, a relevant goal is to achieve an efficient but also effective supply chain through the continuous improvement of logistics processes. This may increase operational costs but also supports the readiness of weaponry within acceptable costs [5,33]. This is a conflict-cost situation in a typical SCM trade-off decision.
Consequently, the defense sector has adopted business SCM practices, such as managing inventory levels [33], providing visibility and aligning appropriate information across the supply chain, improving cooperative and collaborative efforts to enhance the supply chain [31], and transferring some tasks to more efficient private companies via outsourcing [34]. These practices, besides reducing costs, may affect the risk level and vulnerability within the supply chain and even its adaptability and resilience capabilities.
However, "importing" those SCM best practices from the private to the defense sector is not easy, as there are differences between military logistics and supply chain and its private sector counterpart. These differences are striking in scope, scale, and environment. No private or public organization manages and controls a vast supply system and an assortment of supplies, equipment, and personnel as the defense forces. These supplies need to reach many combat "customers" (processing millions of requests) that may differ in their requirements and location [6,35].
The scale of military logistic operations refers to the enormous quantity of items and weight that must be mobilized, far superior to any regular business operation [35]. Moreover, private businesses operate in a relatively predictable and generally peaceful environment. By contrast, military supply chains and logistics must operate in an environment that is often hostile, highly uncertain, volatile, complex, and ambiguous [6,35].
Military logistics and supply chain operations are essential to any national armed force. The support during an ongoing deployment, the adequate response to any threats or natural disasters, and dealing with adverse conditions are critical in reaching operational efficiency. To accomplish these multiple purposes, the defense government institution must foster an improved, resilient, and adaptable supply chain by identifying its level of vulnerability and mitigating the supply chain risks to potential supply chain disruptions (terrorism, cyberattacks, unreliable suppliers, natural disasters, labor strikes), and trying to minimizing the cost and damages [33,36].
A perfect example of an unexpected supply chain disruption is the COVID-19 pandemic that the world is experiencing, which has affected the supply chain of countless products and services in all sectors. For the defense sector, the adverse effects of this pandemic are evident in medical supplies [37], food [38], military equipment, and various products and services [39], not only due to the cascading effects derived from the economic crises that have affected supply chains by volatility in both supply and demand [40, 41] but also because countries were forced to reduce national defense budgets to address the social and public health effects of the pandemic [42,43].
The COVID-19 outbreak has exposed several risks and vulnerabilities, which has compelled nations and organizations to develop adaptability and resilience in their supply chains [44,45].

Resilience
Disruptions in the supply chain have become a critical issue for nations and organizations in today's unpredictable and fluctuating environment, especially troubled by the coronavirus pandemic. Global sourcing and outsourcing of relevant activities raise the risk of supply chain disruptions, making it challenging for managers to build resilience in their supply chains by identifying, managing, and mitigating such risks.
Although resilience is a multidisciplinary concept, in the Supply Chain Risk Management (SCRM) field, it is considered an adaptive organizational capacity that allows the supply chain to prepare, respond and recover from any interruption, often unexpected and in challenging conditions, maintaining operational control of its functions at an acceptable level [46,47]. Following the disruption, resiliency also allows any network to return to its previous state or adapt to a new (and better) operating state [48,49].
There is no specific blueprint in the literature on achieving resilience in the supply chain. However, five organizational capabilities are considered crucial in developing resilience: 1. Re-engineering: All supply chain members must understand their tasks and involvement and be aligned if a disruption could arise. 3. Agility and adaptability: Agility comprises the flexibility and ability to respond adequately to environmental influences through the continuous search for a better response to change and uncertainty (short-term), facilitating the processes of coordination [50,53,54]. Adaptability is the organizational ability to restructure the supply chain, adjusting it to the environment over time by sensing market changes, threats, and opportunities (long-term) [55][56][57].
4. Risk awareness: There must be a clear and shared awareness among the members of the supply chain about their vulnerabilities and latent and potential risks so that, through joint support, they can evaluate them and join efforts to mitigate them [47, 50,52].
5. Knowledge management practices: It is necessary to develop the ability to learn from past disruptive events to better prepare for future events, which can be achieved by sharing what is learned with customers, suppliers, and employees about risks in the supply chain, thus obtaining greater awareness and a better understanding of its structure and the information needed to act [47, 50,58].
In pandemic times, supply chain resilience must focus on generating real-time adjustments while achieving proactive management in designing flexible redundancies that minimize external uncertainty in the supply networks [59]. Recovery and learning emerge as capacities that promote organizational resilience: the first allows resuming the activities before the interruption. In contrast, the second allows improving such activities as a result of what was learned during that interruption and its results [60].
The scope of supply chain resilience may be extended beyond survival by enabling intertwined supply networks to assimilate and endure an event as shocking as the COVID-19 pandemic [61,62]. Other resilience strategies that may help to withstand such events are reallocating resources, restructuring the supply chain, developing more vital collaboration among its members [63], and increasing productivity by reducing costs through digitization (cloud-based services, AI, blockchain, big data) and automation, as resources allow [64], which may help to reduce both external and internal risks.

Risk
The term "risk" describes unexpected events that affect the objectives. Within organizations, decision-makers associate risk with the threat of something that may happen and may disrupt the continuous flow of activities or prevent reaching the planned goals [65].
Risk refers to the chance (probabilistic measure) that a specific threat happens and the consequences (measure) of that event. Any event can occur inside (e.g., financial default of a supplier) or outside (e.g., an earthquake destroys production capacity) of the organization. Consequently, the organization must understand and manage the risk adequately and include in their business strategy their level of exposure to uncertainties and the potential consequences [65][66][67].
Although there is no unanimous consensus, five sources of risk related to the supply chain can be identified in the literature: environment, demand, supply, processes, and control. Environmental risks are those uncertainties that come from the political/legal, socio-economic, or natural environments that surround organizations. Demand risks are associated with the logistical flow of products to the market, while supply risks are associated with supplier relationships. Both (supply and demand) are generally internal to the chain. The design and implementation of flows and processes within and between supply chain members are considered another source of risk, as is the control of management agreements or policies established between such members. It is necessary to note that the sources mentioned can affect each other since environmental risks can lead to supply or demand risks in the chain, while the processes and control mechanisms, in turn, may amplify or absorb the effect of certain external or internal risks in the supply chain [68,69].
The typhoon in Southeast Asia, the terrorist attacks in the USA in 2001, the SARS outbreak in 2002, the tsunami in Japan in 2011, and now the COVID-19 pandemic have revealed severe weaknesses, vulnerabilities, and risks in supply chains and their management [70,71]. Consequently, risk management has become an essential factor in the organizational strategy and in designing SCRM processes, which implies a coordinated effort between the organizations involved to identify and manage risks, mitigate short-term and long-term vulnerabilities, and trace an operational effort toward the future [69]. Accordingly, the first hypothesis contends that risk identification has a predictive influence on vulnerability level (H1).
However, regardless of the risk management strategy, supply chain risks should be identified and managed holistically for an integrated supply chain [72]. In a permanently changing environment, SCRM may be an organizational ability to be flexible and adaptable [73]. Therefore, the second hypothesis asserts that risk identification has a predictive influence on organizational adaptability (H2).
Additionally, organizations must identify risks and respond appropriately to them, regardless of the risks' sources. A better understanding of such sources not only improves the operational performance of the supply chain but can also help to better prepare it for disruptions by adjusting logistics and other activities [74]. Therefore, the third hypothesis in this research states that risk identification has a predictive influence on supply chain resilience capability (H3).

Vulnerability
The term "risk" is also frequently used to describe something vulnerable or susceptible to damage [75]. Organizations must implement a proactive SCRM to identify their vulnerabilities since disruptions adversely affect supply chain outcomes (performance or financial) [76,77]. The features of a supply chain are critical for determining its vulnerability and impact, including both the likelihood of and damages caused by disruptions [66].
Supply chain vulnerability includes the supply chain characteristics, variables, infrastructure, related processes, and environmental factors [78,79]. Vulnerability is visible when multiple events present a risk to the supply chain that could cause unexpected interruptions in the flow of information, products, and knowledge along the chain [80]. The coronavirus pandemic had a profound impact on the global financial and economic order by demonstrating that global supply chains and their distribution structures are highly vulnerable to such disruptions, which will undoubtedly lead to a structural reconfiguration of previously globalized supply chains [81,82].
Consequently, defense organizations must handle the supply chains' vulnerabilities and risks to avoid potential interruptions and consequences, recognize the dynamic nature and extent of supply chain risk, and understand that supply chains are inter-organizational networks operating in an unpredictable environment. Nevertheless, it remains the managers' obligation to identify, manage, and mitigate the effects of risks to reduce vulnerabilities in the supply chain [83] through developing organizational capabilities such as resilience and adaptability to face those uncontrollable forces. Accordingly, the following hypotheses argue that there is a predictive influence of vulnerability level on resilience capability (H4) and adaptability (H5).

Adaptability
According to [84], an organization gains a competitive advantage when its supply chain is agile; it can react rapidly to unplanned changes in demand or supply, it can adapt over time as market composition changes, and it can align the members' interests in the supply network, optimizing the supply chain's performance and maximizing their interests. However, defense organizations do not require a competitive advantage but the ability to answer agilely to different challenges [85].
An organization develops adaptability when it has the skills to adapt to the environment to meet changing customer needs. The supply chain must evolve and change in tandem with the customers. As a result, supply chain adaptability trains its members to respond to evolving circumstances [84,86], similar to the resilience concept discussed earlier. Accordingly, the following hypothesis proposes that there is a predictive influence of adaptability on resilience capability (H6).
However, it is necessary to establish the difference between agility and adaptability in an organizational context. The ability to adjust quickly to short-term market changes, such as variations in demand or supply patterns, shortages, or disruptions, is referred to as supply chain agility [84,87]. The ability to rearrange the supply chain design and respond more radically to long-term developments is known as supply chain adaptability [55,84]. The organizational capability to respond to changes can be characterized as a dynamic capability, a higher-order capability, which promotes the ability to detect opportunities and challenges and capitalize on them by shifting assets and organizational structures [56]. Adaptability is a dynamic transforming capability because the change in the supply chain structure is radical in the long term [55].
The COVID-19 outbreak exposed severe vulnerabilities in the supply chains of uncounted industries, including defense organizations, due to the use of the "benefits" that globalization allowed (before the pandemic) and the natural interdependence between the organizations involved, regardless of their geographic location [44,82]. Part of the adjustments that are seen coming to the industry includes the fundamental restructuring of those global supply chains, which requires adaptability to mitigate threats and discover new opportunities in the long term [2, 81].

Methodology
Based on the literature review, the selected factors affect the supply chain resilience capability, but all the factors also affect each other. Thus, the primary goal of this study is to build on and expand existing research, as well as to present a theoretical model that analyzes the proposed hypothesis (see Fig 1): The purpose of this research is confirmatory-correlational. A confirmatory analysis explains and quantifies the relationships between variables (Yin, 2013). Based on a thorough examination of the literature, a self-administered online questionnaire was constructed according to the parameters established by Hair et al. [88] and Evans and Mathur [89] guidelines. To avoid the social desirability bias, the questionnaire was reviewed by the ethics committee at the Colegio de Estudios Superiores de Administración (CESA) in Colombia to ensure that questions were neutral, unbiased, and non-threatening. Additionally, before filling out the questionnaire, the respondents were informed that all information provided would be treated in the strictest confidence, that the responses would be aggregated and used for research purposes, and that by completing this survey, they were giving their consent to such use, and that they might withdraw at any time. Consequently, consent was given by continuing with the survey, and at any point, the respondents had the possibility to close the online survey. The researchers followed ethical procedures during this research project by adherence to ethical standards.
The survey's instrument was sent via email to Colombian Air Force employees from different areas linked to the institution's supply chain, which has several facilities and a vast network of suppliers. This study based its inferences on analyzing the behaviors of the mentioned factors in the selected institution at two different times: Survey 1 was sent a few months before the coronavirus outbreak, and data collected was analyzed in January 2020, and Survey 2 was sent during the pandemic and analyzed in August 2020.
To assess the face and content validity, we extensively reviewed the literature to determine the relevant variables for the research. Subsequently, we performed a pilot with subject matter experts to define the final instrument. The surveys contain a demographic section, followed by a conceptualized set of variables to build a structural equation model for testing them using descriptive and inferential statistical analysis of the information collected. A Likert-type scale with five points was used to score statements relating to the operationalization of the model's variables, with 1 representing strongly disagree and 5 meaning strongly agree. We added the survey as S1 Appendix. From Survey 1 (before the coronavirus outbreak), 120 questionnaires were returned that were considered valid and usable for the analysis, while from Survey 2 (during the pandemic), 216 were returned that were considered valid.
Information analysis was carried out using SEM. SEM is useful for testing hypotheses concerning relations between latent and observable variables by simultaneously estimating a set of multiple regression equations [90,91]. Furthermore, SEM allows analysis of several dependent relations that capture the impacts of possible mediating constructs [91].
A confirmatory factor analysis (CFA) was performed to examine the relations between the latent and observable variables and determine the model's robustness, validity, and reliability; a confirmatory factor analysis (CFA) was performed. To avoid cross-loadings and latent constructs being correlated, the load values (loading factors) were estimated, and it was validated that each item was used only in one construct. The Cronbach's Alpha coefficients and the overall correlation of items were used to assess the model's internal consistency [91,92].
Reviewing the demographic data, Fig 2 indicates the different educational backgrounds of the respondents, which shows a healthy diversity among them. Fig 3 shows the number of suppliers, companies, or third-party providers with whom there is a relationship in the respondent's area of work. In contrast, Fig 4 shows the number of people working in the supply chain or related activities in the respondent's work area. In both cases, none of them exceed 50%, which indicates the diversity of the sample. Fig 5 shows that most of the respondents (around 90%) take some risk prevention training in their work area; however, the remaining portion does not receive any, which can be highlighted as an opportunity to improve and promote risk prevention training for all employees of the institution. Table 1 shows the values of Cronbach's Alpha coefficients of the constructs. All the coefficients have values above 0.7, established as the minimum acceptable cut-off level for confirmatory research [93][94][95]. Table 2 shows the values of chi-square to degrees of freedom (CMIN/DF) of 1.695 and 1.915, which, being in a range between 1 and 5 [96] or in a range of 2:1 or 3:1 [97], can be taken as indicators of the robustness of the hypothesized model [98,99]. Table 3 shows the robustness measure baseline comparisons, where all values are greater than 0.7 and close to or above 0.9, which are the recommended values [100], as well as the CFI (Comparative Fit Index) and GFI (Goodness-Of-Fit) in Table 4, also considered acceptable in each model since they are around the value of 0.9 [99,101].
The RMSEA (Root Mean Square Error of Approximation), which is acceptable as it is less than 0.08 [101][102][103], is shown in Table 5.
The comparisons reviewed to measure robustness, reliability, and stability suggest that the model proposed is well represented by the variables and related data. Therefore, it is possible to make inferences based on these results.

Compared results
The results shown in Table 6, and Figs 6 and 7 allow inferring that the identification and management of risk have a strong and positive impact on vulnerability reduction in the supply chain, which is supported by the theory [69], thus confirming hypothesis H1 both before and  during the COVID-19 pandemic. At the same time, risk identification and management positively impact the organizational capability to be agile and adaptable [73], thus confirming hypothesis H2 before the pandemic and moderately confirming hypothesis H2 during the pandemic. It is likely that the extent of risk prevention training, even if it only reaches around 90%, indicates the people's awareness to achieve this result [104,105].
However, the results suggest that identification and management of risk do not guarantee the development of resilience capacity in defense institutions since hypothesis H3 was not confirmed before or during the COVID-19 pandemic. This could be explained by theory since, in addition to adapting, resilience capacity implies that the supply chain prepares itself for unplanned events, responds to disruptions, and recovers from them by keeping running operations at an acceptable level of control [47], which is not so easy to achieve simply through training; it requires a more significant institutional effort and commitment [106,107].
Results also show that hypothesis H4 was moderately confirmed before the COVID-19 pandemic and confirmed during the pandemic, suggesting that a decrease in vulnerability in the supply chain positively influences its resilience capability, which is supported by the literature [83].
An interesting result is shown when analyzing hypotheses H5 and H6 that were confirmed during the pandemic, which suggests that the decrease in vulnerability has a strong and

PLOS ONE
Supply chain comparative Covid-19 study positive influence on the improvement of adaptability, which, in turn, has a strong and positive influence on supply chain resilience acting as a moderating variable between vulnerability and resilience. Both results are supported by the literature [83,84]. However, since hypotheses H5 and H6 were not confirmed before the pandemic, it may be suggested that adaptability, such as the organizational ability to radically restructure supply chain design to adjust to long-term market (environment) changes [55,84], requires a significant institutional effort commitment [106,107], and this ability to adapt was not a priority before the pandemic as the supply chain was operating well. It may be suggested that the significant impact caused by the pandemic on supply chains forced organizations to be aware of their risks and vulnerabilities [44,45], forced them to adapt and, consequently, to develop their resilience capacity, in a much shorter term than usually is presumed to be reasonable [59,62,108].

Discussion and conclusion
Christopher [109] noted that supply chain interdependency becomes more widespread when complexity and risk levels rise. Therefore, identifying and managing risks positively impact reducing vulnerability in the supply chain and its capability to adapt to unexpected events that may cause disruptions in its continuous flow. Most respondents had received some training in risk prevention, which undoubtedly contributed to this result. However, there is an opportunity to broaden the scope of risk identification and management training to the people across the entire institution, not only those involved in supply chain activities.
Five categories of supply chain risk sources have been identified in the literature: environment, demand, supply, process, and control [68,69]. For specific risk sources affecting some

PLOS ONE
level of the organization's supply chain, adequate classification and identification are valuable to take more customized mitigation and control measures beyond taking general measures of prevention or reaction because simple identification and management of risks do not guarantee the development of resilience capacity in institutions, as the results showed. To this end, an institutional commitment is suggested that goes beyond adaptability capacity, supporting programs that enable the supply chain to prepare as well as possible, respond adequately, and recover quickly to sudden disruptions while retaining a desirable level of control and continuity of operations, which strengthens resilience.
As this document is being written, the world is facing one of the worst global crises of the last seven decades because of the COVID-19 outbreak, causing severe disruptions in the supply chains of several industries due to their global interconnectivity. The pandemic exposed vulnerabilities in supply chains that were previously considered adequate. Institutional commitment mentioned above must also accompany continuous improvement processes in building resilience capability involving the people in charge. Such processes must also be institutionalized so the same process feeds back to itself, understanding resilience as a capability to be developed constantly.
Finally, it is recommended that further studies be performed to validate the effects of the COVID-19 pandemic on the supply chain of institutions included in this study, to evaluate the level of preparedness, response, and recovery they possessed to face the event, to measure the damage caused, and, more importantly, to synthesize the lessons learned. This will assist in developing plans for containment and reaction to even less future severe events.

Limitations
The research was conducted within the Colombian Air Force, part of the country's defense organization. The Colombian Airforce is highly engineered; consequently, it remains to be seen if the results can be extended to other branches of the defense force, namely the Navy and Army.