Not All Energy Agents Maximize Profits: Modelling Complexity of Investment in Oil Gas Projects

摘要

Researchers from the Brandeis International Business School,in collaboration with King Abdullah Petroleum Studies and Research Center(KAPSARC)are building an agent-based simulation model that encompasses the scenario building efforts of Shell,et.al.on energy and climate models(Shell,BP,IEA,EIA,IPCC).This research contributes to the literature on the complexity of economic systems,showing how heterogeneity,path dependency,feedback loops,and learning by agents can dramatically alter price expectations and outcomes – investment and production--predicted by traditional(equilibrium)economic or VAR models(e.g.,how NOC or IOC agents react to price changes/surprises,shale production,and climate change goals/rules).Analyzing and parsing the investment and production decisions of different producers gives us the tools to model the dynamics of energy markets.We look at agents'endogenous investment behavior to gain a better understanding of investment cycles and to model the transition from higher cost fields(regions)to a more sustainable future.The underlying premise of our approach is that an agent-based model that uses a flexible structure can simulate market interactions and more particularly explain the investment and production cycles.In other words,energy producers have different investment/profit maximizing functions and the heterogeneity of investment matters.The investment,production,and cash flow actions of National Oil Companies,Independent Oil Companies and Shale producers,operating in fields with different costs affects energy supply and,of course,prices.Our agent based,fuzzy logic model lets us to run"what if"simulations by changing common language assumptions(e.g.behavior rule: invest more in shale if prices are high/over $60 a barrel;expand low cost oil & gas fields if expected demand/price peaks in five years).By using field level data to estimate agent investment functions derived from heterogeneous profit expectations we explain the differences of oil production of individual agents and resulting market dynamics.We use an agent model to analyze how shale oil and traditional producers enter/exit the oil market(invest)in response to price expectations,volatility,and cycles.In this paper,we are parsing out the investment behavior and actions of different production agents investing in different fields/projects throughout the world.It is our premise that a simple profit maximizing function does not explain the varying investment decisions of energy agents.