Presentation on theme: "Indian Food Supply Chains: A Game and Model to study Economic Behavior Sebastiaan Meijer, Jayanth Raghothama, Robin King, Bharath M Palavalli Next Generation."— Presentation transcript:
Indian Food Supply Chains: A Game and Model to study Economic Behavior Sebastiaan Meijer, Jayanth Raghothama, Robin King, Bharath M Palavalli Next Generation Infrastructure Lab, Center for Study of Science, Technology and Policy
Agenda Model of semi-perishable food chains, based on real data Initial validations and applications Discussion of applicability of such methods in the context of developing countries
Indian Agriculture & Horticulture 60% of India’s population subsist on agriculture Less than 15% of India’s income 20% price inflation on Agricultural products during the past year Inflation despite buffer stocks and exports and growing production India accounts for 11% of world production in vegetables, and 15% of fruits Total production volumes of around 230 Million Tonnes India accounts for only 2% of global exports Only 2% produce processed Nearly 30% wasted ~ 60 Million tonnes Data Sources: World Bank; Ministry of Agriculture, Government of India
Mango: An Example India accounts for 40% of world production, 25% of exports Over 300 varieties of Mangos grown; differentiated by price, use, location and quality (and taste) Prices range from Rs 8/kg to Rs 300
Mango: An Example Investigate the questions of waste, power structures and inefficiencies in the supply chain using Mango as an example Surveys in “mandis” (markets) around Bangalore, in Karnataka, interviews with farmers and other stakeholders Mapping of supply chain, and data collection to feed into an agent based model, and later a game Model based on surveys and supply/demand data sourced from Govt. of India
Problems within the supply chain APMC (Agricultural Produce Marketing Committee) Act: Mandates a government licensed intermediary through whom products have to be sold Created a layer of Middle-men Middle men provide services in a bundle to farmers ahead of harvest, creating lock-in and skewed power structures High logistical and intermediary costs, credit constraints, inadequate infrastructures, incomplete information etc.
Mango Mandi Gaming Simulation Agent Model can run independent of the game. Players in the game compete and negotiate with software agents.
Attributes Agents located in five cities, four cities who produce and one who only consumes mangos Game lasts for 45 days, in compressed time Inventory every day to all agents based on supply/demand data Mangos: Ripeness, type, price All attributes updated every day
Agent Preferences Agent TypeMango Attributes (ordered, weighted) Scrooge Factor (greed, 0 to 1) History (preference for previous agents, 0 to 1) Transport Type Dealer1.Ripeness 2.Type 3.Price Very High Equal preference for all types Exporter1.Type 2.Price ModerateHighNone Small Vendor1.Type 2.Price 3.Ripeness Very LowLowNone Super Market1.Type 2.Ripeness 3.Price Moderate to High LowNone FarmerVery LowModerate to High Cheapest option preferred.
Validation Game is a work in progress Structure of the underlying ABM based on primary data collected in “mandis” around Bangalore Processes used by agents derived from interviews and various government regulations Varied supply chain structures mean it can be validated only within a particular context
Applications Institutional economics experiments: to determine factors that drive actions in the supply chain Agent-human interaction: study of how effectively agents can interact with human players and vice-versa Understand actor behavior to better design interventions in the supply chain
Applicability Severe lack of data Data when available is not disaggregated, or consistent Context too varied to arrive at a generalized problem definition and structure Can you “translate” a model from one context to another?
Acknowledgements: Jamsetji Tata Trust, India and Next Generation Infrastructures Foundation, The Netherlands Thank you