COMPUTATIONAL ECONOMICS

The PhD Program in Economics offers advanced training in economics, econometrics and finance and aims to provide highly selected students with the ability to conduct original research, enabling them to pursue a career in academia or in economic and financial organizations. This course aims at strengthening quantitative and computational skills exploring comprehensive modelling issues, agent-based models and their computational application.

a) Knowledge and understanding:
- know terminology and concepts related to modelling issues;
- know terminology and ideas to describe agent-based models;
- know terminology and main features of NetLogo programming.

b) Applying knowledge and understanding:
- ability to setup a simple model and its main goals;
- ability to formally define an agent-based model fitting the given purpose;
- ability to use, re-use or develop some computational draft of the model using NetLogo.

c) Making judgements:
- understand and critically assess the merits and problems of any modelling effort;
- make sense of the structure and internal consitency of an agent-based models, suggesting, if appropriate, revision and further invetigations;
- write reports/papers on models and results, positioning the content in the literature and assessing the novelty and importance of the outcome.

Students should be familiar with the material taught in their first year of the PhD program. Knowledge of a programming language is helpful (however, NetLogo is typically quite different from standard tools and, hence, don’t despair if you have no coding experience!).

Tentatively, the following topics/material will be covered:
- several fundamental Agent-Based Models are presented and we discuss how they shed light on interesting emerging phenomena in complex economic, financial, and social systems. We will introduce the Shelling Segregation Model, the El Farol Problem (Minority game), the Santa Fe artificial stock exchange model, and cooperation models. These models reveal important idea related to inductive thinking and economic outcomes, emergent phenomena, and agents' self-organization.
- programming in Netlogo. Working versions of most the previously listed models will be analyzed and studied, many models described in the textbook by Railsback and Grimm will be coded from scratch and modified/generalized. If time permits, an epidemiological SIRD model will be used as a programming testbed.
- ideas taken from complexity theory and genetic algorithms (e.g., related to variation, interaction and selection).

[RG] Railsback and Grimm, “A Course in Individual-based and Agent-based Modeling”, Princeton University Press, 2012.
[A] Ashlock, “Evolutionary Computation for Modeling and Optimization”, Springer, 2006.
[AC] Axelrod and Cohen, “Harnessing Complexity”, Basic Books, 2000.
[N] Netlogo website and documentation, see http://ccl.northwestern.edu/netlogo/ The files will be downloaded to your PC when you install Netlogo.

In the past, grading was based on:
- three presentations in Pechakucha style, see http://www.pechakucha.org/ to be done in couples (week 1, 2 and 5): 30%
- one individual working paper describing an agent-based model and related NetLogo code (expected at the end of week 4): 30%
- final written or oral or practical exam: 30%
- active participation in class and online discussions: 10%

More information will be given at the beginning of the course or whenever possible.

Lectures, practical PC sessions, presentations. Prof. Righi will present ABMs in 5 meeting (once per week), prof. Pellizzari will mainly teach NetLogo programming in the remaining 10 meetings.

The course will present well-known and insightful agent-based models taken from economics and social sciences, as well as fundamental ideas and computational methods to develop your own models. The student will also become familiar with NetLogo, an object-oriented programming language, coding platform and state-of-the-art tool in computational economics (download the code at http://ccl.northwestern.edu/netlogo/ ). The student will reach the level needed to select his/her modelling ideas, implement an agent-based model in NetLogo, verify and document the code using the ODD protocol and improve the presentation skills needed to demonstrate the functioning of the computational model and its scientific use.

written and oral