Computational Methods in International Political Economy: Simulations, Network Analysis, and Machine Learning

The integration of computational methods into International Political Economy (IPE) represents a transformative advancement, enhancing both the depth and breadth of research in the field. These methods use algorithms and simulations to study large or complex settings. enabling scholars to connect micro-level interactions with macro-level outcomes, addressing concerns about the discipline’s focus on micro-foundations. In this entry, we detail three main techniques: To start, we describe simulations through agent-based modeling, where the behavior of agents (e.g. individuals, companies, or states) and their interactions is defined by a mathematical model and run on a computer hundreds of times. This exercise allows for an understanding of how the interactions of individual agents (e.g., diplomatic relations between countries) lead to larger systemic changes, such as the impact of tariff policies on global trade. Next, we detail network analysis, where relationships between agents are represented and visualized as networks composed of nodes (agents) and ties (connections between them). This allows IPE researchers to, for example, map trade relations, and identify which countries hold the most bargaining power. Finally, we discuss machine learning, where computational algorithms are used to identify patterns and relationships in the data. This approach enables scholars to process large volumes of data (e.g., millions of text documents), thus allowing them to develop and test theories using novel datasets. These three relatively accessible tools—simluations, network analysis and machine learning—provide powerful means for analysing complex data sets, offering new insights into global economic dynamics and informing more effective policy-making.