Five theses on neuroeconomics

Abstract

Over the last decade, neuroeconomic research has attracted increasing attention by economic modellers and methodologists. In this paper, I examine five issues about neuroeconomic modelling and methodology that have recently been subject to considerable controversy. For each issue, I explicate and appraise prominent neuroeconomists' findings, focusing on those that are claimed to directly inform economic theorizing. Moreover, I assess often-made assertions concerning how neuroeconomic research putatively advances the economic modelling of choice. In doing so, I combine review and critical arguments to provide a methodological evaluation of neuroeconomists' contributions.

Keywords:

• neuroeconomics
• choice models
• economic theory
• interdisciplinary integration
• modelling pluralism

Acknowledgements

I thank J. McKenzie Alexander, Richard Bradley, John Davis, Wade Hands, Don Ross and Jack Vromen for their comments on earlier versions of this paper. I also benefited from the observations of two anonymous referees and audiences at the University of Münster, the London School of Economics and the Erasmus University of Rotterdam.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

^1. NEs frequently use the terms ‘models’ and ‘theory’ interchangeably when presenting and discussing their findings. In this paper, I follow this terminological practice unless the cogency of my considerations hinges on specific differences between these two notions (see, e.g. Hausman, 1992, chapter 5; Mäki, 1996, section 5).

^2. I expand in Section 3 on the challenges faced by NEs' calls to employ neural findings for model selection purposes in economics. As to the predictive and explanatory benefits yielded by NEs, I articulated in previous works two challenges to NEs' calls to use neural findings in constructing and evaluating economic models. The first challenge (Fumagalli, 2011) is that due to the trade-offs between the desiderata respectively valued by NEs and other economists, showing that neural findings help economists to satisfy specific desiderata falls short of implying that NEs enable economists to build better models of choice. The second challenge (Fumagalli, 2014) is that NEs have hitherto failed to significantly improve economic models even with regard to individual desiderata.

^3. I speak of ‘neural substrates of choice’ broadly to indicate both algorithmic and neuro-anatomical findings. The proponents of NE frequently emphasize that NE studies target both sets of findings. For instance, Montague argues that there are ‘two natural [NE]’, one which investigates ‘the behavioural algorithms running on [the] neural tissue’, the other which examines ‘the way that neural tissue is built, sustains itself through time, and processes information efficiently’ (2007, p. 219).

^4. Expressions such as ‘informing economic models’ and ‘informing the economic modelling of choice’ are used in different senses by distinct authors. In this paper, I employ such expressions to refer to the thesis that NE findings have direct evidential and explanatory relevance for the economic modelling of choice.

^5. Fumagalli (2010) occasionally contrasts incremental and radical NEs, as opposed to incremental and radical NE. Here I focus on the latter contrast, since several authors provide both incremental and radical contributions in their studies and are therefore hard to categorize as either incremental or radical NEs.

^6. Neural findings are claimed to serve discriminatory purposes in various disciplines besides economics (see, e.g. Kable, 2011, on the opportunity to combine brain-imaging and brain-stimulation findings to discriminate between psychological theories). Here I focus on the use of such findings for model selection in economics. In doing so, I employ the expression ‘available evidence’ broadly to indicate the set of behavioural, psychological, neural, etc. findings that are taken to bear on the merits of the examined models unless I mention a specific subset of those findings.

^7. Some authors (e.g. Ross et al., 2008, p. viii) claim that standard economic models do not rest on explicit assumptions as to what sort of entities (e.g. human individuals, firms and neural areas) representative agents supposedly map onto. Yet, even those authors grant that economic modellers are ultimately concerned with living organisms of some sort (see, e.g. Ross, 2009; see also Davis, 2009, for a discussion).

^8. To give one example, Rosenberg infers that economics is a biological science on the sole ground that economists study ‘the causes and effects of the behavior of members of [the Homo sapiens] species’ (2009, p. 60). The observation that economists investigate the causes and effects of humans' behaviour does not per se make economics a biological science in an interesting sense. Moreover, Rosenberg puts forward questionable claims in explicating the implications of the alleged biological character of economics. In particular, he nowhere substantiates his claim that ‘almost everything mysterious and problematical […] about economics is resolved once we understand economics as a biological science’ (2009, p. 59).

^9. NE has been claimed to foster revolutionary changes in other respects besides the one examined in this section. For instance, Glimcher and Rustichini (2004) emphasize the scope of NEs' proposed revolution, which supposedly encompasses the accounts of choice respectively developed by economists, psychologists and neuroscientists. For his part, Glimcher (2011, chapter 6) insists on the depth of NEs' purported revolution, which allegedly fosters the reduction of some basic constructs employed in NE's parent disciplines. I gloss over these considerations here, since I already commented on them in Sections 1 and 4.

^10. This does not imply that NE is plausibly regarded as the mere continuation of behavioural and experimental economics with technologically more sophisticated instruments. I am not concerned here with discussing how NE differs from earlier research programmes at the interface between economics and psychology. For present purposes, it suffices to note that NEs frequently criticize former works in behavioural and experimental economics. For instance, Glimcher et al. allege that most bounded rationality models ‘have little or no predictive power outside of their bounded domains’ (2005, p. 214), and Glimcher contends that prospect theory ‘has too many interacting parameters to [be regarded as] a truly falsifiable theory’ (2011, p. 120).