Risk management: The intersection of supply chain and aircraft availability

From the March-April 2025 edition of Supply Chain Management Review.

March-April 2025

Inside this month's issue of Supply Chain Management Review, we look at the complicated process of managing parts for military aircraft and what private sector supply chain managers can learn. Plus, understanding…

Browse this issue archive.

Access your online digital edition.

Download a PDF file of the March-April 2025 issue.

Supply chains encompass a range of activities that stretch from raw material sourcing to the delivery of an end item for customer use. They are often complex global ecosystems that require stakeholders to employ supply chain risk management (SCRM) practices that can mitigate disruptive events. The national security implications of exposed Department of Defense (DoD) supply chains make them a prime candidate for employment of SCRM practices; identifying risk factors that may be used to calculate the vulnerability profile of individualized elements of the supply chain. Preemptive identification of risk factors is pivotal to mitigating disruptive impacts arising from supply chain distress.

While supply chains are often simplified as three-tier ecosystems, the structure can extend to multiple layers or subcontracted activities required to source raw materials, manufacture, and deliver an end product. Management of the parts associated with a weapon system (referred to as NIINs or National Item Identification Numbers) may be spread over several stakeholders who, in turn, engage with suppliers at the tier-one level.  The notional supply chain illustrated in Figure 1 is useful when framing the potential vulnerabilities inherent to a military supply chain. The intricate web of suppliers stretching from tier one through tier three forms the activities required to mine and shape raw materials into the requisite specifications for installation onto the applicable weapon system. This complex network fosters an environment where risk factors may complicate the enterprise’s task of developing a cohesive understanding of the supply chain vulnerabilities.

 

The White House 2022 National Security Strategy (NSS) utilizes the phrase supply chain in 19 unique instances, calling attention to “insecure supply chains,” “supply chain security,” and “protecting supply chains” from political pressure and potential national security threats from countries like Russia and the People’s Republic of China (PRC). The supply chain prevalence in the NSS signifies a clear national security pillar. Developing a DoD SCRM model to assign individual risk scores to individual part numbers (NIINs) of a major weapon system may be a small, but impactful step to assess the national security risk. By calculating and mapping impact and probability scores onto a risk matrix, synthesizing data from various sources, analyzing NIIN lists, detailed NIIN data fields, governmental reports, and other multi-agency contributions, units can uncover and address vulnerabilities at the NIIN level, providing actionable insights for which components require collective resourcing from supply chain stakeholders.

To guide the study and dig deeper into the collected data, this research focused on a list of parts within a single military aircraft, analyzing risk factors to assign a risk level to a given set of components, as well as how these risk factors can be combined to create a comprehensive tool for calculating a supply chain risk score. A comprehensive risk management model was developed to enhance the alignment of enterprise logistics efforts. The model was designed to pinpoint NIINs with elevated risk and offer critical insight into potential areas where future constraints to aircraft availability may emerge, while offering key NIINs for potential resource investment.

Data collection

The primary data collection method was solicitation to military, governmental, and commercial stakeholders with supply chain management roles for the studied aircraft.  To narrow the scope of research, one aircraft was focused on for this study and the NIINs contained in an iteration of the 2023 MRSP were the subject of analysis. MRSPs are pre-assembled, air transportable, kits of critical aircraft parts and related maintenance supplies necessary to maintain aircraft in the field for a set period of time. Three organizational stakeholders were identified as responsible for at least one of the 285 NIINs selected. These three organizations provided data for this study in conjunction with a commercial consultant organization.

Additional quantitative and qualitative data was pulled from various military, governmental, and commercial reports accessed via unclassified systems. These reports introduced information ranging from the state of military supply chains, supply chain risk factors, geopolitical considerations related to raw materials, and other relevant areas.

Risk factor identification

The Supply Chain Backorder Risk Trigger (SCBORT) model outlined by Rodger et al. (2014) is uniquely applicable given the study included military supply chains, risk management, and NIIN evaluation. Figure 2 demonstrates the similarities between the SCBORT model and the aggregated risk factors from the other included studies. Though all studies use different nomenclatures, all sampled works included individualized risk factors that captured (1) quality, (2) financial implications, (3) time, (4) scale of production or service, and in the case of the SCBORT model (5) the availability of information (Kabak & Oztek, 2022; Rodger et al., 2014).  The similarities between these studies gave credence to the SCBORT model and informed the structure of this research. The SCBORT approach provides a qualitative description of each trigger dimension. 

 

This study distilled the five SCBORT trigger dimensions into weighted impact and probability measurement metrics, as shown in Table 1. Risk factors within this model are selected based on their severity and likelihood of representing a disruptive condition relevant to aircraft availability (AA). All metrics used for impact and probability calculations are normalized and weighted on a 0 to 1 scale to ensure no risk factors disproportionately influence final risk scores. 

The number of aircraft Mission Impaired Capability Awaiting Parts incidents (MICAPs) represent the risk factor for the Impact-Quality category. A MICAP is generated when a component failure either fully or partially reduces the capability of the aircraft. MICAP events directly constrain AA due to parts or maintenance-related issues. The raw number of MICAPs directly hindering AA was interpreted as a measurement of quality. The Probability-Quality measurement within the model captured the rate of MICAP occurrence relative to all demands for the component.

 

The Impact-Cost category risk factor is the current NIIN price per individual unit.  Many of the NIINs associated with an MRSP are considered repairable items. The new purchase price of the given NIIN may not be reflective of hidden costs such as activating a new production line or inflation. The Probability-Cost metric for the cost category was the number of the given NIINs required for each aircraft. NIIN requirements per aircraft were determined to be a method for gauging the likelihood of a requirement arising to purchase the given NIIN.

The Impact-Lead Time risk factor utilized net repair hours as the metric to represent the amount of time for a functional part to be generated by the supply chain. All NIINs within the sample are considered repairable, making net repair hours a suitable substitute for acquisition and production lead time. The Probability-Lead Time risk factor within the model was the NIIN failure rate, which is specific to aircraft flight hours. This measurement captured the average number of flight hours before a NIIN failed onboard the aircraft and was interpreted as a measure of the likelihood that the repair lead time process could manifest as an impact. Higher values for failure rate indicate greater risk. 

The Impact-Service Level risk factor is measured by calculating the hours required to replace all MICAPs for the given NIIN. This measurement was selected as a representation of impact as it represented a level of support needed to replace the NIIN and to return the aircraft to full mission-capable status. The Probability-Service Level risk factor was the Mean Time Between Failure (MTBF) metric. The MTBF fundamentally represents the total operational time in hours that the component is installed on the aircraft before failing. Higher values for MTBF indicate lower risk, as the component has a longer operating life expectancy. This metric differs from NIIN failure rate, as MTBF includes both ground and air operations. 

The Impact-Information Availability risk factor was represented as the number of Product Quality Deficiency Reports (PQDRs) for the given NIIN over the component’s history. The Probability-Information Availability risk factor measures the rate of PQDRs by measuring the number of aircraft PQDRs within a 24-month period against the number of PQDRs over the component’s history. The PQDR process has been implemented to capture various disruptions that may arise through design, development, purchase, production, supply, maintenance, and other functions.

Creating a risk matrix

Impact and probability weights were derived from the literature (Rodger et al., 2014). The SCBORT model developers scored the quality of a NIIN as being the risk factor with the greatest likelihood of impacting operations—in this case, the aircraft availability (AA) of the studied aircraft; shown in Table 2. Moreover, the likelihood of the quality risk factor arising was the highest of all five risk factors. Lead time scored the lowest impact among the risk factors but the second highest when considering the probability of disruption affecting operations. Information availability scored the second highest among the risk factor impacts, but the lowest for probability of the risk factor impact operations. 

 

These weights were transformed to create an Air Force risk matrix utilizing a 5-point scale for probability measurements and a 4-point scale for impact measurements. The impact and probability scales were directly carried over from the official Air Force regulation, Department of the Air Force Pamphlet (DAFPAM) 90-803, Risk Management Guidelines and Tools. All impact and probability measurements were normalized onto the appropriate scale and graphed using the Air Force risk matrix template.

As shown in Figure 3, this study adapted a risk matrix based on products from the United States Air Force and academia. When populated, the matrix visually represents the relative probability and impact risk levels to which each NIIN is vulnerable. The matrix adapts the SCBORT aggregate risk levels to the Air Force categories of low (L), medium (M), high (H), and extremely high (EH). Such a matrix is valuable for identifying risk clusters within a sample of NIINs.

 

Employing the risk matrix

Using enterprise-standard risk factors, risk scores were calculated over two tests: (1) all NIINs with academia-informed weights, and (2) top and bottom 10 NIINs with academia-informed weights. Both tests visualized results using an Air Force risk matrix. The model’s results are relative to the sample of NIINs scored in a given iteration. All variables were normalized onto a 4-point scale for impact, and a 5-point scale for probability that allowed NIINs to be graphed onto the sample risk matrix shown in Figure 3. Impact and probability totals for each NIIN were weighted according to Table 2 and subsequently graphed based on the finalized scores.

Test 1

The first test scored all 285 NIINs using the weights shown in Table 2 (scores summarized in Table 3). Individual NIINs were graphed on the template matrix in Figure 4 depicting a large concentration of NIINs in the matrix’s low-risk region. The results suggested that 262 NIINs (92%) fell into the low-risk area of the matrix, and 23 (8%) were scored as medium-risk.

 

A summarized heat map of impact and probability scores for all 285 NIINs is shown in Table 3. Data analysis suggests that MICAP-related variables were influential to the risk profile. This is perhaps no surprise, as MICAP conditions either partially or fully constrain the operational capability of the aircraft, and the model is oriented to AA disruptions and heavily weighted toward MICAP triggers. NIINs 142 and 78 were two of the highest scores when averaging impact and probability levels from this model. These high scores were anticipated, as the NIINs had high values across all impact and probability variables compared to the rest of the sample size. 

NIINs 160 and 231 scored among the lowest impact and probability scores. All 10 risk factors for these NIINs were measured at the low end in relation to the population of 285 NIINs. No MICAPs were registered during the 24-month window for either NIIN, with minimal MICAP hours for the service level category. Moreover, these NIINs registered as low-cost NIINs compared to the full sample of 285 NIINs, with NIIN 160 registering a cost of $3,322 and NIIN 231 costing $515. The component with the highest associated price was NIIN 45 with a listed price in excess of $915,000.

 Probability scores were more varied in their variable predictors. NIIN 142 scored high in probability, with the lowest MTBF value and average inputs in all other categories. NIIN 117 registered an average MTBF, but the highest NIIN failure rate within the sample of 285 NIINs. This suggested that while NIINs 117 and 142 had among the highest likelihood of disruptive factors occurring, the associated impact of these factors was not definitively linked with elevated impact levels. NIIN 142 was assigned an impact score of 1.986, whereas NIIN 117 was assigned an impact score of 1.096. Three impact categories separated the two NIINs where NIIN 142 scored higher in comparison to NIIN 117: (1) Number of MICAPs, (2) NIIN Cost, and (3) MICAP Hours. 

Test 2

Impact and probability scores output by this SCRM model are relative to the sample of included NIINs. As such, a grouping of NIINs most and least likely to constrain AA may be assigned a different risk level when included as subgroups. This phenomenon is expected as constraining NIINs are anticipated to have similar risk factors within the model. As the model leverages the process of normalization, similar data elements are likely to suggest the absence of elevated risk. Alternatively, if these constraining NIINs are run against components less likely to impact AA, then the model is reasonably expected to assign less concentrated impact and probability scores. Test two scored the top 10 and bottom 10 scoring NIINs from test one, using the weights shown in Table 2. The results of this test showed that of the 20 NIINs, 11 (55%) fell into the low-risk area of the matrix, and nine NIINs (45%) were scored as medium-risk. Figure 5 graphs these 20 NIINs onto the Air Force risk matrix.

 

Table 4 shows a heat map of impact and probability scores comparing the model output of 285 NIIN risk scores and the subset of 20 NIINs. This table demonstrates the impact of running a non-random subset of NIINs against the full sample size of 285 NIINs. A noteworthy result of this test is that probability scores averaged higher, contributing to a 25% increase in the medium-risk category being assigned to a given NIIN compared to the full sample results.

 

The majority of the NIINs received similar impact scores between test one and two. However, 17 of the 20 NIINs (85%) saw an increase in probability scores, and three NIINs (15%) were assigned lower probability scores than in test one. Impact scores were not perceptibly changed at three decimal places. This phenomenon was likely attributable to the limited data variance between top and bottom-scoring NIINs within the impact category, whereas there was considerably more variance within the probability category. 

The set of 20 NIINs was further broken down into two derivative tests where the top and bottom 10 scoring NIINs were scored separately. Figure 6 graphs the 10 NIINs with the highest aggregate risk score, and Figure 7 graphs the 10 NIINs with the lowest scores.

 

 

Results from Figure 5 compared to Figures 6 and 7 generally demonstrated the localization effect of the model. The top 10 scoring NIINs consistently held high values in both impact and probability categories. This appeared to result in minor changes to overall scoring when the results of Figures 5 and 6 were compared. The inverse appeared true when the results of Figures 5 and 7 were compared. Figure 5 shows the bottom 10 scoring NIINs as a cluster in the low-risk zone. When scored separately in Figure 7, the distribution of these same NIINs stretched into medium and high-risk zones of the risk matrix.

Conclusion

Several stakeholders in DoD and academia have highlighted the challenges inherent to SCRM. For military logisticians, challenges are exasperated by obsolescence, a diminishing supplier base, vast networks of subcontracted suppliers, and even the lack of a consensus in the SCRM framework. This model shows promise in assigning NIIN-level risk scores when aggregating several performance metrics to aid in data-driven decisions on resource investment decisions. By using a range of variables, stakeholders are provided with a diversified view of risk that can be manipulated through a number of supply chain resilience efforts.

Ultimately, this research serves as a bridge from qualitative SCRM approaches to a quantitative SCRM model, yielding an objective risk score based on key commodity metrics. The study also outlines a framework for visualizing large numbers of commodities on a single risk matrix in order to pinpoint factors that may be influencing impact and probability factors contributing to supply chain issues that manifest as constraints to aircraft availability.    

Research priorities for the DoD supply chains have been identified through several sources and range from raw material sourcing, sole-source suppliers, supplier visibility, and obsolescence. Quantitative approaches are of greater value to stakeholders than qualitative ones, as quantitative modeling may allow organizations across the DoD and industry to rally behind consensus risk indicators. Subsequently, these qualitative models may allow for multiple organizations to be statistically represented within the SCRM, thereby enhancing the risk score’s credibility. 

It is in any organization’s best interests to consider methods of properly allocating finite resources. An enterprise portfolio of several million unique commodities is not conducive to reliably targeting NIINs with the greatest risk predictors relative to aircraft availability, and perceptions of risk may differ throughout the supply chain as a customer’s view differs from the suppliers. Therefore, the incorporation of a SCRM model that leverages agreed-upon data elements to calculate a risk score is a valuable tool when seeking to create a decision advantage on which parts to resource, for which aircraft, and with what sense of urgency. In both DoD and industry, supply chain performance plays a crucial role in the organization’s ability to reach and exceed its performance metrics; therefore, it is imperative that investments in the supply chain are prioritized based on the overall risk to the supply chain’s ability to support those metrics. 

References

Kabak, M., & Oztek, G. (2022). A Multi-Criteria Approach to Sustainable Risk Management of Supplier Portfolio: A Case Study at Defense Industry. Gazi University Journal of Science, 35(4), 1504–1519. https://doi.org/10.35378/gujs.883894

Rodger, J. A., Pankaj, P., & Gonzalez, S. P. (2014). Decision making using a fuzzy induced linguistic ordered weighted averaging approach for evaluating risk in a supply chain. International Journal of Advanced Manufacturing Technology, 70(1–4), 711–723. https://doi.org/10.1007/s00170-013-5311-4

About the authors

Rob Delaney is Logistics Readiness Officer in the United States Air Force with extensive experience in operational logistics, including assignments with the Defense Logistics Agency and NATO. He can be reached at [email protected].

Dr. Amanda Femano  is an assistant professor of Logistics and Supply Chain Management at the Air Force Institute of Technology. A Logistics Readiness Officer in the United States
Air Force, she is a trusted logistics leader dedicated to tackling operational problems with data-driven decisions. Dr. Femano’s research interests include supply chain resilience, transportation, and strategic mobility. She can be reached at [email protected].

SC
MR