Academic journal article Erasmus Journal for Philosophy and Economics

Behind the Model: A Constructive Critique of Economic Modelling

Academic journal article Erasmus Journal for Philosophy and Economics

Behind the Model: A Constructive Critique of Economic Modelling

Article excerpt

Review of Peter Spiegler's Behind the model: a constructive critique of economic modelling. Cambridge: Cambridge University Press, 2015, 201pp.

Peter Spiegler examines the current state of theoretical and empirical modelling in economics. According to him, this involves answering two questions: First, how can one determine what causal factors are essential for a phenomenon and, second, how does one ensure that the methodological tools employed represent these features faithfully (p. 9)? Spiegler puts forward new answers to these questions. He argues that a look behind mathematical models in economics is necessary: these methods are themselves not capable of showing that the economic phenomena and their representations within formal models are compatible, in the sense that the formal models conceptualise their domain of applicability accurately for a given epistemic purpose. He suggests that this compatibility between formal methods and economic phenomena needs to be checked with the help of an interpretativehermeneutic method, akin to techniques used in anthropology and sociology, and that this should give rise to a new subfield of economics: interpretative economics.

Economists and philosophers of economics alike might be puzzled by this suggestion of giving qualitative methods such a key role within economics. I think both should be stimulated by Spiegler's proposal. Throughout the book, it becomes evident that he is a philosophically highly informed economist who identifies relevant issues in a precise manner and skilfully navigates through the nitty gritty details of particular episodes of economic modelling. Even if one leaves aside his call for a substantial reform of economics, this book contains a lot of food for thought. For example, his discussion of the New, New Institutional Economics and Dynamic Stochastic General Equilibrium (DSGE) modelling provides rich case studies that put on the radar of philosophers of economics subfields of the discipline that have so far not received enough attention.

Spiegler's two specific claims are the following: Blind spots are a problem for all formal methods in economics and there needs to be an interpretative-hermeneutic method to assess the aptness of these formal techniques. I am intrigued by these two claims but not convinced by them. My worry is that Spiegler does not do enough to support them. Before spelling this out, let me briefly summarise the three parts of the book.

In part I, Spiegler introduces a pragmatic account of formal modelling and puts forward a criticism of theoretical and empirical formal modelling that sets the stage for the rest of the book. He suggests the following framework that is inspired by Mäki (2009): An epistemic agent S uses a model M to represent A for purpose P. The success of S in accomplishing P is judged against disciplinary norms N (p. 25). He further differentiates between four stages of formal theoretical modelling: In the delimitation phase, a social phenomenon is delimited and a research question is formulated in ordinary language (e.g., why is there involuntary unemployment?). In the denotation phase, the delimited social phenomenon is connected to a mathematical model in two steps: first, the formal structure of the model is described informally-with the help of ordinary language names for the phenomenon (yielding what he calls a proto-model); second, the model is presented in purely formal terms. In the solution phase, purely mathematical operations are performed to arrive at a result. In the interpretation phase, the solution stated in mathematical terms gets retranslated into ordinary language using the correspondence established in the denotation phase (pp. 46-52). According to Spiegler, the same four phases can be used to describe econometric modelling. In this case, however, two additional relations need to be accounted for: 1) econometric models are (sometimes) models of an economic theory and 2) econometric data is data about economic phenomena (p. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.