Some Alternatives to Model Selection and New Approaches to Computing for the Economics. 12/6/2014

No Thumbnail Available

Meeting name

Sponsors

Date

Journal Title

Format

Thesis

Subject

Research Projects

Organizational Units

Journal Issue

Abstract

This interdisciplinary dissertation in statistics, economics, and social science methodology derives a methodological approach for economics from philosophical principles, identi es barriers to its adoption, and presents tools and methods for overcoming those barriers. It addresses aspects of inquiry often treated in the discipline of statistics. The dissertation takes a process philosophy approach encompassing aspects of classical pragmatism and documents the in uence of this problem solving philosophy on Original Institutionalism and its support for a link between Post Keynesian economics and Original Institutionalism. I motivate my contributions with two case studies. The statistical methods used for policy persuasion in the rst case study, in community economics, contrast with the econometric approach and its philosophical and theoretical foundations. After a survey of the Keynes-Tinbergen and Cowles Commission-NBER debates, I conclude that economists should shift emphasis including favoring data analysis over statistical inference, and argue that the methodological approach to statistics built on the problem solving philosophy reclaims the tradition of Keynes and the original institutionalist and joins it with best practice in modern statistics. The second case study, in modern statistical theory focusing on alternatives to model selection, reveals the importance of computational thinking and the institutional adjustments the discipline of statistics carried out in embracing that. I epistemologically justify a new computing paradigm developed from this case study. I then detail a ceremonial role of mathematical rhetoric in the intransigence of economic methodology. I compare and contrast the mathematical rhetoric of Keynes and Samuelson and trace intellectual roots for an instrumental approach toward mathematical rhetoric that join Keynes to the Original Institutionalists through process philosopher and mathematician Alfred North Whitehead in the United Kingdom and Charles Sanders Peirce in the United States. I close with a Moore Method inspired course based on this work which provides a practical start for revising the economics curriculum to reclaim the Keynesian- Institutionalist tradition addressing emerging needs in the discipline as a whole and to train students in the methodological approach I have here worked out.

Table of Contents

Introduction -- Philosophical background -- Problem solving philosophy -- Case study in community economics: lawns to street signs -- Statistical practice in economics 75 years after the Keynes Tinbergen debate -- Case study in modern statistics: adaptive regression by mixing (ARM) -- Computational thinking -- Overcoming the obscurantism of the mathematical rhetoric of economics -- Conclusions -- Appendix A. Comparing model combining methods

DOI

PubMed ID

Degree

Ph.D.

Rights

License