Formal Approaches to Socio-economic Analysis—Past and Perspectives

Abstract

This paper is motivated by the observation that (1) socio-economic analysis uses significantly less formalisms than mainstream economics and (2) that there exist numerous situations in which socio-economics could benefit from a more formal analysis. This is particularly the case when institutions play an important role in the system to be investigated. Starting with a broad conception of formalism, this paper introduces and discusses five different formal approaches regarding their adequateness for socio-economic analysis: The Social Fabric Matrix Approach, the Institutional Analysis and Development Framework, System Dynamics, (Evolutionary) Game Theory and Agent Based Computational Modelling. Every formalism entails implicit ontological and epistemological tendencies that have to be reflected on if the formalism should contribute to a better understanding of the system under investigation. The above-mentioned formalisms are no exception. Therefore, this paper pays particular attention to these tendencies. In the end, antagonisms and possible convergences among the formalisms are discussed.

Keywords::

• social economics
• institutional economics
• methodology
• epistemology
• ontology
• system dynamics
• social fabric matrix
• (evolutionary) game theory
• agent-based computational economics
• econometrics

JEL classifications::

• B41
• B52
• C63

Acknowledgements

The author wants to thank Wolfram Elsner and Henning Schwardt for their valuable comments on an earlier manuscript. All the remaining errors are my own.

Disclosure statement

No potential conflict of interest was reported by the authors.

Grand-Awarding bodies:

There were no grant-awarding bodies involved.

Notes

¹ I have extended this framework and illustrated its usefulness for the examples of agent-based models and dynamic stochastic general equilibrium models in Gräbner (2015).

² The social fabric matrix and the institutional analysis and development framework are examples of formalisms in this broader sense.

³ Note that there are other formalisms that have been successfully applied by socio-economists. Yet, it is impossible to consider all of them in this single paper. The selection was made in accordance with my impression about the overall relevance of the formalisms, the degree of controversy associated with their application and with my personal preference. For an overview on other candidates see Radzicki (2003), Lee and Cronin, (2015) or the special issue of the American Journal of Economics & Sociology (Lee, 2011).

⁴ For a corresponding compilation see Hayden (2006a), or consider Tool (2003) and Fullwiler et al. (2009) for a summary of the impact the SFM-A had on public policy.

⁵ A more thorough introduction can be found in Ostrom (2005).

⁶ Attributes of a community include, but are not limited to, the level of trust and reciprocity, habits, the value structure or cultural dispositions.

⁷ A thorough introduction can be found in Sterman (2000).

⁸ Note that Fiddaman (2002) has not considered any cultural habits of values explicitly in his model. But these could be added into the model without great difficulty.

⁹ For a very good introduction into game theory and evolutionary game theory from an institutionalist perspective, see Elsner et al. (2014).

¹⁰ Strategies are often interpreted as genes, but also as values, behaviour, habits or the like. There exist settings in which the strategies themselves are under ongoing change and players develop new strategies, according to the rules of Darwinian evolution, see e.g. Lindgren (1992).

¹¹ ABMs have been used in the social sciences since the 1990s, with Epstein and Axtell (1996) and their “Generative Social Science” as a major vantage point. But there were some predecessors in the 1960s, particularly in the field of cellular automata theory.

¹² Note that the program does not tell you to which ontological level the different software objects belong. This information must be given by the theory underlying the model.

¹³ If one considers the economic agents to represent the micro level, ABMs are suited for microcalibration. This involves a direct test of the adequateness of the agent design. It is common to consult field experts to judge the behavioural assumptions or to exchange the agent with a real human being “playing” the role of the agent in the model and then by comparing how the software and the real human being have behaved. Because the behavioral specification of the agents is done via computer code, there is no upper limit for the complexity of the rules other than accountability considerations. Chen (2012) describes various types of agents including very elaborated artificial intelligence agents.

¹⁴ Unfortunately, most model applications do not reflect explicitly on their epistemological orientation, which makes it difficult to assess the models from a critical perspective.

¹⁵ The extensive survey of Chen (2012) shows how different types of agents can be implemented within the ABM framework. The range of possibilities starts with zero-intelligence agents that may not be of primary interest for social economists, but also include elaborate artificial intelligence agents that require a lot of effort to be built. The adequate conception depends on the problem at hand. The important lesson for social economists is that a very detailed description of agency is possible.

¹⁶ For Ostrom, examples of theories are game theory, public choice theory, transaction cost theory and the like. I define theory in a broader sense such as the neoclassical or institutionalist theory, but this is not of crucial importance. I agree that the theory suggests the parts of the framework that are given particular importance.

¹⁷ This shows that a clear-cut distinction between the three ideal types is not always possible. Also, different researchers apply the formalism in a different manner. Still, they may be useful as a general taxonomy which facilitates to think about formalisms and their application.

¹⁸ Hodgson (2013a) rightly points to the confusion that is often associated with individualism when he discusses the meaning of methodological individualism. Whether individualism includes the relation among individuals has never been clarified. I think that the constituent element of individualism in any sense is the denial of downward effects on individuals. This is in line with Kapeller (2015), who identifies this denial with a simple fallacy of aggregation.

¹⁹ Some passages of Marx suggest a clear holist ontology, e.g. his statement that “the capitalist functions only as personified capital […] just as the worker is no more than labour personified.” (Marx, 1982, p. 989). But there are also other examples, e.g. the notion that “history is nothing else than persons pursuing their aims.” (Marx & Engels, 1956, p. 125).

²⁰ In fact, theory is required to allocate the different objects of the ABM on different ontological levels. Technically, an agent is as much an object as an institution. That the latter belongs to a different ontological level must be inferred from the underlying theory.

²¹ Scholl (2001) argues for a certain complementarity among SD and ABM: They give insights to the same system from different perspectives. This does not affect the argument that they are not to be used simultaneously together.