PyData London 2017: Forecasting social inequality using agent-based modelling

James gave a brilliant talk about forecasting social inequality using ABM at the recent PyData London 2017 conference. You can check out the video here and the slides here.

The abstract:

“How can we assess the future impact of changes to government policy? One tool that is gaining in popularity is agent-based modelling, in which a population of agents – typically representing individual people – is simulated together with an environment they can interact with. The agents’ actions, which can in turn influence the other agents and the environment, are governed by a predetermined set of rules or heuristics. After much use in fields such as ecology and epidemiology these models are gaining increasing recognition in economics and policy, and were recently featured in the Bank of England’s Quarterly Bulletin. By running parallel simulations with modifications to the environment or the behavioural rules, a policy-maker can compare the likely outcomes of different options available to them.

We have used an agent-based model to assess the outcome of a change to the way in which wealth is inherited within families. The proposed change favours putting wealth into trust funds for grandchildren and great-grandchildren, instead of passing it directly to children. A similar proposal covering only houses was made last year by Gavin Barwell, the housing minister, but was not taken forward as government policy.

We developed a simulation of the demographic makeup of the UK, based on data from the census and the ONS, and inserted the different inheritance methods into it. We could then see what the long-term outcome of each scenario would be, in terms of the distribution of wealth and level of inequality in the country, allowing a quantitative assessment of its impact on individuals and society as a whole. The model also provides a base for developing assessments of more complex policies and interventions.”

Comments are closed.