The Non-Evolutionary and Non-Benign Character of Stylized Facts

Abstract

This article highlights the potential of stylized facts to take on a non-evolutionary and non-benign character. Two stylized facts are used as primary examples throughout the article: the inverse trade-off between changes in unemployment and changes in price level (i.e., the Phillips curve) and that countries with debt-to-GDP ratios in excess of 90 percent experience lower economic growth. The insights of Pierre Bourdieu will be drawn upon to understand the socialization processes which create common sense understandings of our world that are non-evolutionary and non-benign, what Bourdieu calls doxic understandings. Doxic understandings of the world are inherently ceremonial and antithetical to an evolutionary approach. To overcome these problems, does not mean the elimination of heuristics; but rather, a continual application of reflexivity. Reflexivity is a process, where the researcher is continually interrogating what is being taken for granted in their own methods and logic, to ensure these are grounded in instrumental reasoning. Reflexivity is necessary to safeguard economics and the public against doxic understandings, and is a necessary condition of realism in analysis.

Keywords:

• stylized facts
• performativity
• Phillips curve
• Reinhart-Rogoff
• reflexivity
• doxa

JEL Classification Codes:

• A2
• B4
• B5
• E6

Notes

1 First identified by Phillips (1958) through inductive methods, for a specific time and place.

2 Associated with claims of Reinhart and Rogoff (2009, 2010).

3 See Snowdon and Vane 2005, 305–307 for an example.

4 Lawson characterizes realism as, “realism asserts the existence of the objects of research as independent of the enquiry of which they are the objects. In other words, according to this doctrine, there is a material and social world that exist independently of any individual consciousness and which is knowable by consciousness—true theories of real entities can be obtained” (1989, 61).

5 This is why this article is titled “The Non-Evolutionary and Non-Benign Character of Stylized Facts,” and not titled, “The Non-Evolutionary and Non-Benign Nature of Stylized Facts.” In this case, “character” is used to imply characteristics taken on through deployment as opposed to “nature” implying something intrinsic, irrespective of deployment methods.

6 As pointed out by Deirdre McCloskey, in lieu of conversations regarding the epistemological questions of thresholds regarding model misspecification or an empirical relationship no longer holding, we are left in this infinite, “pseudo-scientific ceremony of hypothesis-regression-test-publish” (1983, 499).

7 For an example of effective performativity, see Donald Mackenzie (2007) discussing the Black-Scholes model and its influence on the process generating option prices.

8 You might be tempted to argue that the Phillips curve is not assumed, as evidenced by the number of empirical studies each year investigating it. However, these studies, when not finding an inverse relationship often fall back to questions of model misspecification rather than asking if the relationship has broken down. This tendency of researchers is emblematic of the assumed nature it has taken on in their habits of thought.

9 For example, see the Winter 2018 edition of the Review of Keynesian Economics wholly dedicated to such an interrogation.

Additional information

Notes on contributors

Jacob Powell

Jacob Powell is an assistant professor of economics at Hobart and William Smith Colleges.