A Plea for (Good) Simulations: Nudging Economics Toward an Experimental Science

Abstract

In this article, the author argues that simulation is an undervalued technique to draw conclusions about empirical phenomena in economics. If the aim is to learn about the behavior of socioeconomic systems of interest, simulations have a variety of advantages relative to alternatives such as mathematical (pen and paper) modeling and laboratory experimentation. Therefore, the author has a good prima facie reason to exploit this method more fully. The author proceed by demonstrating that frequently heard arguments against simulations are wrong, and finally the author discusses a number of more specific empirical phenomena, criticisms of one type of simulation methodology used in economics.

Simulation-based research in management accounting and control: an illustrative overview

“…First of all, research based on simulation is often regarded as suffering from what is called a "black box": models and results often suffer from not being comprehensible and transparent to other researchers (Lorscheid et al 2012;Harrison et al 2007;Reiss 2011;Barth 2012): a major virtue of simulation models is allowing to analyze rather complex phenomena (particularly, compared to "traditional" formal models); making simulation models fully comprehensible and transparent to other researchers, often would require extensive descriptions of data, rules, and parameter settings, etc. [which often remains undone to avoid overloading the reader or to meet the publisher's space limitations (Lorscheid et al 2012)].…”

Computer simulation in exprimental economics

“…W literaturze przedmiotu prowadzi się od lat szeroką dyskusję na ten temat (Guala, 2002;Morgan, 2003;Grossklags, 2007;Parker, 2009) i wykazuje następujące zalety tego podejścia (Reiss, 2011): możliwość precyzyjnego powtórzenia eksperymentu symulacyjnego, możliwość zmiany parametrów (które nie mogą być zmieniane w systemie rzeczywistym), znacznie niższe koszty eksperymentowania (finansowe i etyczne) oraz krótszy czas realiacji eksperymentu.…”

Improving simulation model analysis and communication via design of experiment principles: an example from the simulation-based design of cost accounting systems

“…see Balakrishnan et al 2007;Harrison et al 2007). Creating scenarios is also a benefit in situations in which real experiments are restricted for practical reasons (Reiss 2011). Moreover, simulations support the development of new theories (Davis et al 2007) and experiments can be conducted in a highly controlled environment (Klingert and Meyer 2012;Rubinstein and Kroese 2008).…”