![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
The paper discusses the extent to which the availability of unprecedentedly rich data-sets and the need for new approaches – both epistemological and computational – is an emerging issue for Macroeconomics. By adopting an evolutionary approach, we describe the paradigm shifts experienced in the macroeconomic research field and emphasize that the types of data the macroeconomist has to deal with play an important role in the evolutionary process of the development of the discipline. After introducing the current debate over Big Data in social sciences, the paper presents a detailed discussion of possible and existing interactions between Big Data and Computational Behavioral Macroeconomics. We argue that Big Data applied to economic questions can lead to new styles of thinking and research methods, namely to the development of a new research paradigm.
Acknowledgements
The author acknowledges the participants of the 2016 Duke Forest Conference (Durham, NC, USA), the editors of the Journal of Economic Methodology, prof. John Davis and prof. Wade Hands and an anonymous referee for useful comments and suggestions. The author is also grateful to prof. Michael Roos for comments on an earlier version of this paper. All usual disclaimers apply.
Notes
No potential conflict of interest was reported by the author.
1 According to an online survey of 154 global executives conducted in April 2012 by Gandomi and Haider (Citation2015), the majority of the interviewed executives defined BD as the ‘massive growth of transaction data, including data from customers and the supply chain’; others as ‘new technologies designed to address the volume, variety, and velocity challenged of BD’. The 18% identified them with the ‘explosion of new data sources’ (social media, mobile device, and machine-generated devices) and the 19% focused on the requirement to store and archive data for regulatory and compliance. The reported clustered answers make clear that executives differed a lot in the understanding of the term Big Data, showing also some confusion with respect to what BD are and what they do.
2 By CBM we mean Agent-based computational models ‘informed’ by findings of Behavioral Economics (see Akerlof, Citation2002; Colander, Howitt, Kirman, Leijonhufvud, & Mehrling, Citation2008; Velupillai & Kao, Citation2014, among others).
3 Regarding the use of survey data in macroeconomics (see Bricker, Henriques, Krimmel, & Sabelhaus, Citation2016; Carroll, Citation2003; Dunkelberg, Citation1986; Ormeño & Molnár, Citation2015).
4 Regarding the latter point, BD motivate the development of new computational infrastructure in terms of computational efficiency, data-storage methods and they pose significant challenges also for software development (See Kambatla, Kollias, Kumar, & Grama, Citation2014, for a detailed review on this issue.
5 It was instituted in 1913 and offers a large micro-level data-set of tax revenues. It has been used by Piketty and Saez (Citation2003) to derive historical series of income shares for the top percentiles earners among US households.
6 Machine learning is gaining a lot of attention because caeteris paribus of its ability to deal with large data-sets and the possibility of continuous ‘retraining’ of the algorithm over time as the environment changes. For details about the features and implementation of the most popular algorithms we refer the reader to more specific articles and contributions, such as Goldberg and Holland (Citation1988), Birchenhall (Citation1995), Michalski, Carbonell, and Mitchell (Citation2013).
7 A fitness (or adaptive) landscape is a model that comes from biology where it is used to describe the ‘fitness’ of an agent, or more specifically genotypes, within a particular environment. The dynamics of evolution are represented as a search over a set of possible solutions to a given environmental condition, with the aim to find the optimal strategy that will have the highest elevation on the landscape and receive the highest payoff. The better suited the agent to that environment, the higher its elevation on this fitness landscape will be.
8 Consider, for example, the statement by Blaug: ‘ I will argue that the term “paradigm” ought to be banished from economic literature, unless surrounded by inverted commas.’ (Blaug, Citation1980, p. 149).
9 A similar view is expressed by Robert Lucas: ‘I came to the position that mathematical analysis is not one of many ways of doing economic theory: It is the only way. Economic theory is mathematical analysis. Everything else is just pictures and talk.’ (Lucas, Citation2001, p. 9).
10 German for ‘game of language’. The reference is to the work of the philosopher Ludwig Wittgenstein who developed this concept in his Philosophical Investigations in 1951–1953.
11 This concept has been retaken some years later by Trichet (Citation2010) who argues that in order to overcome the limitations of mainstream macro models, economists should consider the interactions among heterogeneous agents and use behavioral economics and agent-based modeling.
12 It is important to mention that similar issues are discussed also among the mainstream economists. Caeteris paribus, a recent contribution by Reis (Citation2017) suggests that there is actually nothing wrong in current Macroeconomics. He claims that the critiques are off target and that current research is not ‘mindless DSGE modeling filled with ridiculous assumptions and oblivious of data’ (Reis, Citation2017, p. 22).
13 An alternative view has been developed by Lehtinen and Kuorikoski (Citation2007), which argue that since simulations cannot correspond to the perfect model, ‘ economists do not therefore consider them viable candidates for generating theories that enhance economic understanding’.
14 The concept of abduction was introduced by Charles S. Peirce. We refer the reader to Niiniluoto (Citation1999) for a more comprehensive discussion about it in the philosophy of science.
15 Nomo-empirical experiments aim at testing propositions about behavior that are suggested by observed empirical regularities in field data or pilot experiments (Smith, Citation1991).
16 See, The BPP: Using online data for measurement and research, VOX CEPR’s Policy Portal, 26 April 2016, http://voxeu.org/article/billion-prices-project.
17 On the same topic, see also the contribution by Richiardi, Leombruni, Saam, and Sonnessa (Citation2006).