Title: Econometrics and computational economics: an exercise in compatibility

Authors: Todd Feldman; Yi Sun

Addresses: Department of Finance, San Francisco State University, San Francisco, CA 94132, USA. ' HSBC Private Bank, 71 S Wacker Drive, Suite 2700, Chicago, IL 60606, USA

Abstract: This paper tests whether the econometric model of Boswijk et al. (2007) (BHM07) adequately identifies strategy switching behaviour by using computational data from a different model, in particular the model Friedman and Abraham (2009) (FA09). The purpose of using computational data based on an endogenous behavioural mechanism distinct from switching behaviour is to examine whether we can recover the estimates found using S&P 500 data. The results indicate that the estimation results from BHM07 can be partly recovered from the computational data suggesting that BHM07 does adequately identify switching behaviour. However, results also suggest that factors not included BHM07 but included in FA09 may be as or more important in explaining fat tails and excessive volatility.

Keywords: financial markets; agent-based modelling; agent-based systems; multi-agent systems; experimental economics; econometrics; computational economics; strategy switching behaviour; fat tails; excessive volatility.

DOI: 10.1504/IJCEE.2011.043250

International Journal of Computational Economics and Econometrics, 2011 Vol.2 No.2, pp.105 - 114

Published online: 22 Oct 2011 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article