Over the past several decades, sustained growth in milk production has been driven by institutional development, policy support and incremental improvements in farm-level practices. Cooperative procurement systems, technological adoption and marketing arrangements have made producers participate in organized dairy markets, thereby boosting regional production capacity and income stability (Kumar et al., 2021; Sharma and Singh, 2022). India’s emergence as the world’s largest milk producer is the reason for these institutional transformations alongside productivity gains across livestock systems (Rezitis and Kastner, 2021; Zheng et al., 2022). Most of the people perceive milk and value-added dairy products as a primary source of protein in their regular diet in India (Malar et al., 2025). Rising consumer demand and evolving dietary patterns have intensified pressures across procurement, processing and distribution systems, exposing structural vulnerabilities within the supply chain (Govindan and Chaudhuri, 2021; Mangla et al., 2021). Persistent infrastructural gaps-particularly in cold-chain logistics, transportation reliability and processing capacity-continue to affect product quality, delivery performance and market responsiveness across regions (Kamble et al., 2022; Saberi et al., 2021). The dairy supply chain functions as an interconnected system involving farmers, collection centres and market participants with financial and informational flows operating continuously across stages. Dairy farmer can improve their income by understanding the variations and tuning the feeding programme and housing practices for improvement (Mishra et al., 2025). Contemporary research advises that disruptions in such systems rarely remain localized; instead, they propagate through interconnected processes, impacting inventory dynamics, price fluctuations and availability outcomes across markets (Ivanov, 2021; Wieland and Durach, 2021). Current technological advancements have contributed to increasing supply-chain visibility and decision-making capabilities within the dairy sector. The emerging adoption of digital tools such as IoT-based monitoring systems and advanced analytics has triggered stakeholders to track quality indicators, monitor procurement flows and forecast demand patterns more accurately (Huerta-Soto et al., 2024; Serrano et al., 2025). Milk analytics has grown as a tool for finding production variability, monitoring quality performance (Lal and Mehta, 2021; García-Méndez et al., 2024). Ardent scholars debate that purely analytical approaches frequently fail to capture delayed responses, nonlinear interactions and cumulative effects inherent in complex food systems (Azizsafaei et al., 2022; Shahsavari-Pour et al., 2023).
       
System dynamics modeling offers a framework for addressing these flaws by representing stocks, etc., shaping supply-chain behaviour over time (Azizsafaei et al., 2022; Govindan and Chaudhuri, 2021). Such approaches are notably relevant for dairy systems, where seasonal production variability, demand fluctuations and logistical constraints interact actively to impact inventory stability and service outcomes. The current study develops a risk-based analytical design, which integrates milk analytics with system dynamics modeling to evaluate the behaviour of the Indian dairy supply chain. By blending empirical production data with econometric analysis and simulation-based modeling, the study focuses to identify prime risk factors, check their interactions within causal-loop and stock-flow structures and generate insights for decision-making among policymakers, cooperative institutions and private-sector stakeholders (Pandit et al., 2024). The integrated approach chosen in this research offers a more comprehensive understanding of how structural and operational factors jointly shape the evolving dynamics of the Indian dairy ecosystem under conditions of uncertainty and structural transformation.
 
Literature review
 
Recent studies urge that supply-chain performance in the dairy sector relies not only on production efficiency but also on the integration of logistics, information flows and institutional coordination mechanisms (Azizsafaei et al., 2022; Govindan and Chaudhuri, 2021), followed by Ivanov, 2021; Wieland and Durach, 2021) showing the significance of risk management frameworks in dairy supply chains owing to growing uncertainty caused by climate change and transportation disruptions. System-level risks in food supply networks have been shown to propagate through interconnected nodes where disruptions at one stage can induce cascading effects throughout the supply system. In the view of perishable commodities such as milk, delays in transportation or inadequate storage infrastructure can lead to significant spoilage losses and reduced product availability in markets. Consequently, effective supply-chain risk management strategies require coordinated actions across procurement, processing and distribution stages (Mangla et al., 2021; Saberi et al., 2021). Moreover, lack of infrastructure leads to higher transaction costs and ineffective supply-chain responsiveness to supply and demand dynamics (Kumar et al., 2021; Sharma and Singh, 2022). In addition, empirical evidence also suggests that areas with better cold-chain infrastructure have more stable operations and less wastage than those with poor logistics infrastructure (Rezitis and Kastner, 2021; Zheng et al., 2022). Information platforms allow real-time tracking of production, stock and quality across the supply chain, demonstrating the use of smart technologies, such as blockchain, Internet-of-Things (IoT) and artificial intelligence, to enhance transparency, traceability and coordination in supply chains (Kamble et al., 2022; Serrano et al., 2025). For example, blockchain technologies enable stakeholders to securely record and transmit transaction information, enhancing confidence and minimising security risk in food supply chains. Likewise, IoT technologies support temperature monitoring during transit and storage, allowing detection of quality issues (Huerta-Soto et al., 2024; Sivasankaran et al., 2023). Data analytics help stakeholders track production levels, predict demand and streamline inventory management. Research on dairy markets reveals how predictive analytics can support procurement strategies and decrease risks in supply-chain management (Lal and Mehta, 2021; García-Méndez et al., 2024). Through the analysis of past production and demand indicators, stakeholders can anticipate emerging supply-consumption mismatches and hence plan logistics and processing activities accordingly (Lyngkhoi et al., 2022). Dairy supply chains are affected by feedback loops, time delays and nonlinear relationships between various factors (e.g., production variability, transportation delays, demand variability, inventory adjustments) (Ivanov and Dolgui, 2020; Govindan and Chaudhuri, 2021). Consequently, there is a growing interest in using system dynamics models to study these interactions. This enables researchers to simulate the effect of operational decisions on system performance over time, using stocks, flows, feedback loops and time delays. Research using system dynamics in agri-food supply chains shows that simulation models can be used to pinpoint areas of intervention to enhance supply-chain resilience and minimize vulnerabilities (Azizsafaei et al., 2022; Shahsavari-Pour et al., 2023). Using empirical data to build simulation models can help assess the effectiveness of policy measures such as infrastructure investment, transportation infrastructure or additional processing capacity. In dairy supply chains, resilience is related to infrastructure preparedness, information openness and coordination between supply-chain stakeholders (Govindan and Chaudhuri, 2021; Wieland and Durach, 2021). Research on sustainable dairy supply chains also highlights the importance of holistic planning that takes into account environmental sustainability, resource efficiency and economic sustainability (Mangla et al., 2021; Saberi et al., 2021). Moreover, researchers have started to explore the role of technological innovation in facilitating sustainable dairy supply chains. Technologies such as artificial intelligence, machine learning and predictive analytics are being applied to enhance animal monitoring, feed efficiency and milk yield predictions (García-Méndez et al., 2024; Serrano et al., 2025). In summary, recent studies highlight the need for holistic dairy supply-chain management. While the current research offers insights into dairy supply-chain management, there are still some gaps in the literature. First, research typically focuses on isolated aspects of supply-chain performance, such as logistics efficiency, infrastructure improvement or digitalization, without addressing the holistic and dynamic impacts on supply-chain stability. Second, empirical research often adopts static analytical approaches that fail to consider feedback loops and dynamic interdependencies in complex supply networks. Third, there is a lack of research combining empirical production analytics with system dynamics modeling in emerging markets, especially in the Indian dairy industry. This current research seeks to address this by proposing an integrated approach that combines empirical data analytics and econometric modeling with forecasting and system dynamics simulation to investigate how risks propagate in the Indian dairy supply chain and strategies for maintaining resilience.
 
Rationale for the study
 
Supply-chain stability in the Indian dairy sector is challenged by structural and coordination challenges among the stages of supply-chain management, which with seasonal variations in producing milk and uncertain demand, results in fluctuations in the balance of supply with consumption, resulting in wastage during the surplus period and scarcity during the shortage period and hence impacts market stability (Govindan and Chaudhuri, 2021 and Ivanov, 2021; Lyngkhoi et al., 2022). Poorer cold-chain networks, logistical delays and inconsistent processing capacities lead to a greater risk of quality and product loss in a highly perishable product category, implying that inefficiencies in logistics and infrastructural imbalances limit the agility of dairy supply chains in emerging markets, especially in regions with fragmented procurement setups and uneven market access (Kumar et al., 2021; Sharma and Singh, 2022). Further, the dairy supply chain is a complex network of farmers, collection points, processors, distributors, retailers and consumers interacting with each other through material, financial and information linkages. Recent research makes it clear that disturbances in such networks are unlikely to remain isolated; instead, they accumulate across the network in feedback loops, impacting inventory, service level and pricing decisions (Ivanov and Dolgui, 2020; Wieland and Durach, 2021). Likely, therefore, static approaches to portray the risk in the food supply chain of perishable products are not sufficient. Rather, we need dynamic analytical models that can represent accumulations (stocks), changes (flows), delays and feedback that influence system dynamics (Azizsafaei et al., 2022; Shahsavari-Pour et al., 2023).
 
Need of the study
 
Despite India’s status as the world’s largest milk producer, procurement, processing and distribution inefficiencies continue to play a key role in the avoidable loss and instability in operations and economic performance (Pandit et al., 2024). Research reveals wastage, fluctuation in farmers’ income and limited market availability are often driven by coordination issues between various supply chain control nodes rather than limited production (Mangla et al., 2021; Saberi et al., 2021). This highlights the importance of integrated analytical frameworks to better represent the interplay between production variability, logistics efficiency and market demand. In this context, milk analytics would help stakeholders to track production and quality indicators and to estimate demand by providing a data-driven platform for decision-making (Lal and Mehta, 2021; García-Méndez et al., 2024) and system dynamics modeling would complement analytics by offering a platform to study aspects of feedback-driven behaviour, adaptation over time and policy interventions within dynamic systems (Azizsafaei et al., 2022; Lyngkhoi et al., 2022; Govindan and Chaudhuri, 2021).

The current research study will have a mixed-analytical design that combines descriptive milk analytics coupled with system dynamics modeling to explore risk behaviour in the Indian dairy supply chain. The analytical framework is so constructed that it would allow the observation of interdependencies, feedbacks and time lags between the system components, rather than just a statistical study. Initially, descriptive techniques were evolved to find production trends, variability in distribution and the structural changes in states. Adopting an official statistical source to obtain a secondary data-stack that illustrates the recent shift in the milk production and per-capita consumption of milk in Indian states.
       
The key time period of analysis is 2014-15 to 2023-24. The duration is enough to capture the cyclical and seasonal changes in the time horizon on one hand and have sufficient time variation to econometrically estimate and argue upon the system-diagramming of value-chains. The whole set of data has been collated using secondary institutional sources such as statistical yearbooks of national dairy development board (NDDB) and other detailed data on livestock sector. At the state level, annual data for milk production were obtained and unified with various measures of measurement. We also prepared per-capita availability data and adjusted these data on the common time periods. Initially milk analytics was carried out using descriptive statistical techniques such as trend analysis, plotting and measuring the growth-rate. However, description was not considered to be sufficient for exploring causal relationships and feedback directed behaviour (Pandit et al., 2024). To address this shortcoming, the authors had calculated a fixed-effects panel regression, to determine the responsiveness of the milk production variables to the indicators of demand and other growth factors and adjusted to the unobserved state-specific effects and general temporal shocks. Further, forecasting has been conducted using ARIMA techniques.

State x year expanded multi-variable panel dataset
 
The empirical analysis has been continued to a balanced state-year panel to overcome the limitations linked to the dependence on the narrow set of national indicators that were conducted between 2014-15 and 2023-24. The statistical tables of NDDB were obtained as the state-wise production and per-capita availability of milk and harmonized across years, bearing in mind the comparability. These were not treated as individual measures but further analytical variables were created to lend more econometric insight and supply-chain arguments. Namely, the dataset will include seven derived measures as shown in Table 1.


 
Data visualizations
 
Visual analysis was done to determine spatial patterns, structural disparity and variation in behaviour among states. Fig 1 shows a state-by-year production heatmap of the top-20 producing states and indicates that the growth trends are clustered in the western and southern parts. Going beyond this first visualization, Fig 2 shows a bubble and connecting production and per-capita availability of 2023-24, the size of the markers indicates national share and the strength of the colours indicates cold-chain efficiency. The fact seems to indicate that there is a significant dispersion in distribution performance among high-output states. Fig 3 also gives additional distributional information showing violin plots of year-on-year growth by quartiles of cold-chain efficiency. Although inconclusive, the patterns of distribution suggest that more volatility is concentrated in the regions of lesser efficiency.






       
This comparative analysis is furthered in Fig 4 with a radar visualization of the top-five producing states in terms of normalized indicators of 2023-24.


       
Structurally, Fig 5 provides a Lorenz curve in which the concentration of production is discussed across states. The curvature shows that there is moderation of inequality as opposed to extreme domination. Lastly, Fig 6 shows system-dynamics scenario results of various performance measures that gives an overall perspective of the dynamic adjustment mechanisms. Fig 7 shows an ACF and PACF of regression residuals, which imply that there is no significant instability even though there is slight serial dependency.
 






Top 10 state production comparison (2023-24): Ranking, shares, CR and HHI
 
The analysis of concentration at the state level shown in Table 2 ensures that there is no concentration of production in a few states that are dominant in production. At the national level, 81.79% of the national output is represented by the top-10 producers (CR10) and 53.99% by the top-five (CR5). The computed Herfindahl-Hirschman Index (HHI = 0.0854) implies the moderate concentration of structure. These indicators suggest that the dominant states have a significant impact on the stability of national supply, but the stability of the system remains based on the wide geographic diversification.


 
Fixed effects panel regression model (State FE + year FE)
 
The panel regression model is a fixed effects panel model that is a combination of State FE and Year FE. A fixed-effects panel regression model was estimated to assess the production responsiveness to different operating conditions with state-level and year-specific controls. The specification of the model was as follows:
 

Milk production_it = β₀ + β₁ per capita_it + β₂ growth rate_it + β₃ demand proxy_it + μ_i + λ_t + ε_it

 
Standard errors were clustered on a state level. The model fit seems also to be good with R² H” 0.991 and the overall F = 10.64 (p<0.001). Although the coefficients of per-capita availability and demand were not statistically significant, the growth-rate variable is clearly significant (p = 0.0025), indicating that adaptive capacity of production is at the center of the responsiveness of the system as presented in Table 3. Further, Diagnostic tests were carried out to assess the reliability of the model. Multicollinearity test shows that the mean of VIF is relatively high (approximately 168.70), indicating the overlap of structural structures between constructed indicators. The test of residual autocorrelation was Ljung-Box (median p 0.031) and Breusch-Pagan tests (p 0.000) were used to test the heteroskedasticity. In estimation, cluster-robust inference was thus held on.


 
ARIMA prognosis (State-level; Leading 5 producers)
 
The top-five manufacturing states were fitted to ARIMA models through AIC-based models’ selection between candidate specifications. The period of the forecast was created to be 2024 25 to 2028 29 as shown in Table-4 displaying the projected mean value of 44.64 (Uttar Pradesh), 37.31 (Rajasthan), 27.25 (Madhya Pradesh), 21.98 (Gujarat) and 19.07 (Maharashtra) and summed up in a consolidated fan chart given in Fig 8.




 
System dynamics scenario simulation (Output curves)
 
Simulations of scenarios were done to test the behaviour of the system under five different conditions that included; baseline operations, larger transport delays (+20%), demand shocks (+15) and processing expansion (+10) and improvement of cold-chain. The results illustrated in Fig 6 indicate that nonlinear growth in spoilage and unmet demand is caused by transport delays using inventory instability processes. Processing expansion stabilizes throughput by absorbing variability in inflow but cold-chain advances produce the most predictable changes in spoilage as well as smoothing fluctuations in inventory.
       
Combined, these findings seem to indicate that infrastructure investments can act as high-leverage interventions that are capable of strengthening stability at interdependent supply-chain steps.

There was continuous growth in milk production and availability per capita in India, indicating a high level of production momentum and increased system capacity. Operational risks such as transportation delay, cold-chain, and inventory will raise the spoilage risks and decline the effective supply to process and distribute. Hence, integrated risk awareness and proactive strategies for supply-chain resilience can be implemented through milk analytics and system dynamics simulation. This study shows that improving supply-chain resilience in dairy systems requires coordinated focus on the above issues to manage seasonality and avoid wastage in periods of excess while avoiding shortages in periods of deficit for effective decision-making, and better regulatory support, infrastructure investments and institutional coordination among stakeholders. Despite the remarkable contributions of this study to the Indian dairy supply chain through the integration of milk analytics, lack of reliance on primary state-level findings, proxy variables increase the difficulty to capture micro-level functional diversity and forecasting accuracy. Hence, the explorations must be interpreted as indicative only, and further research must focus on primary data collection from supply-chain actors, state-of-the-art forecasting and hybrid modeling approaches, comparative studies and development of data-driven and calibrated simulation models to improve accuracy, policy-relentless and resilience in dairy supply chains.

The present study was supported by Co-authors.
 
Disclaimers
 
The views and conclusions expressed in this article are solely those of the authors and do not necessarily represent the views of their affiliated institutions. The authors are responsible for the accuracy and completeness of the information provided, but do not accept any liability for any direct or indirect losses resulting from the use of this content.
 
Informed consent
 
All animal procedures for experiments were approved by the Committee of Experimental Animal care and handling techniques were approved by the University of Animal Care Committee.

The authors declare that there are no conflicts of interest regarding the publication of this article. No funding or sponsorship influenced the design of the study, data collection, analysis, decision to publish, or preparation of the manuscript.