Modelling the economy bottom-up - Professor Doyne Farmer

Professor J. Doyne Farmer is Director of the Complexity Economics programme at the Institute for New Economic Thinking at the Oxford Martin School, Baillie Gifford Professor of Mathematics at the University of Oxford, and an External Professor at the Santa Fe Institute. His current research is in economics, including agent-based modelling, financial instability and technological progress, while his past research includes complex systems, dynamical systems theory, time series analysis and theoretical biology.

Our research focuses on examining the economy as a ‘complex adaptive system’ - a distributed network of dynamically interacting, heterogeneous agents, whose behaviours, strategies and relationships evolve over time. This views the economy as an ecosystem of interacting specialists who have only bounded rationality.

Instead, the economy requires complex system approaches to model it accurately, using fine-grained data sets at the level of firms, products and industries, with the goal of providing a detailed understanding of the rich and heterogeneous behaviour underlying business cycles, inflation and interest rates, innovation, and long-run growth. Agent-based modelling incorporates diverse behavioural models to better capture the dynamic, non-equilibrium aspects of the economy.

This work is being carried out in modules, creating models which will eventually interact.

  • Structure and dynamics of the global production and credit network: We are building models of the global production and credit networks. This involves developing methods to reconstruct missing data, and agent-based modelling of supply chain and trade credit. 
  • Understanding the economic effects of Covid-related lockdown.

During the first wave of Covid-19 in April and May 2020, we developed methods and models to understand the economic effects of lockdown, by evaluating the shocks at the industry and occupation level and by evaluating the effects of these shocks in a non-equilibrium network model of the economy.

  • Heterogeneous dynamics of inflation at a fine-grained level

Inflation rates at the product level are highly heterogeneous and highly persistent. We are developing methods to characterise groups of products or services that have similar patterns of price changes.

  • Forecasting technological progress

Innovation is the dominant factor underpinning economic growth, and is essential for sustainability and the clean energy transition. We have developed time series forecasting methods and network models which we use to map the evolution of technological ecosystems.

  • Network approaches to labour markets and industrial strategies

The labour market is undergoing unprecedented change. On the one hand, new technologies, pandemic restrictions, and climate policies cause employers to demand different tasks from workers, while on the other, the workforce is changing as workers adapt and learn new skills. We use network analysis and agent-based models to analyse how labour demand changes and how workers may better adapt.

  • Green energy transition

We use an empirically validated technology forecasting method based on an expansive historical dataset to model different scenarios for the global energy system. Under a rapid transition scenario energy prices become lower than historical averages after 2030 and considerably lower after 2050, with predicted savings of many trillions of dollars. In contrast, a slower transition is more expensive, while a nuclear scenario is substantially more expensive.

  • Financial stability and systemic risk

The 2008 crisis dramatically demonstrated the dangers of using too much leverage (buying assets with borrowed money), and in particular how this can create systemic risk. We have been working to develop models that can be used to better understand systemic risk and to quantify its dangers.