Money, Banks and Finance in Economic Thought

Text Mining and the History of Economic Thought: Keynes’s Treatises on Probability, Money, and Employment

Komine Atsushi, Ryukoku University

As a preliminary case study of an ambitious project arguing the research relationship between quantity and quality, this study aims at a specific twofold target. First, it will argue whether, in the so-called big data era, it is suitable to apply text analytics to the traditional approaches in the History of Economic Thought (HET), introducing a new quantitative (but more complex) method, namely, text mining. It will argue, particularly, whether the dissection of original text (a whole book, parts, chapters, paragraphs, and sentences) into morphemes (or little pieces), an elementary process of text mining, is compatible with traditional HET methods for interpreting original texts. Second, the study will demonstrate how a leading economist, J. M. Keynes, was consciously or subconsciously consistent throughout his work regarding important themes such as money and uncertainty. More specifically, it will argue to what extent significant elements of his Treatise on Probability (TP, 1921), Treatise on Money (TM, 1930), and General Theory of Employment, Interest, and Money (GT, 1936) are connected with (or disconnected from) one another regarding monetary or uncertainty aspects. Numerical and graphical results will be introduced to illustrate such aspects effectively. Although definite conclusions are still to be finalized, the study will argue at least four aspects. First, applying text analytics to HET is a suitable opportunity to reconsider HET research design in the light of methodological arguments mainly in political science and sociology. Referring to studies such as King, Keohane & Verba (1994), Brady & Collier eds. (2010), and Goertz & Mahoney (2015), this study will deal with the characteristics of hermeneutics, compared with other quantitative or qualitative approaches. Second, it will exemplify both economic studies in digital humanities (social network analysis, topic model, and bibliometrics) and text mining studies in social sciences (social psychology, knowledge engineering, discourse analysis in tourism studies or in the stock market). Third, it will examine the relevance of Keynes’s concept of probability both in the big data era and for his economic thought. He refuted frequency theory in search of a logical and objective relationship of inference in the uncertain world. What does this mean both for Bayesian statistics and for the continuity with his earlier days? Fourth, by using text mining, this study will investigate his three books (TP, TM, and GT) in the light of money and uncertainty.

Area:

Keywords: Text Mining, Keynes, HET research methods, probability, money

Paper file