Applications of Statistical Physics to the Social and Economic Sciences

Petersen, Alexander M. Applications of Statistical Physics to the Social and Economic Sciences. PhD Thesis thesis, Boston University, Graduate School of Arts and Sciences. (2011)

Preview

PDF - Accepted Version Download (2MB) | Preview

Abstract

This thesis applies statistical physics concepts and methods to quantitatively analyze socioeconomic systems. For each system we combine theoretical models and empirical data analysis in order to better understand the real-world system in relation to the complex interactions between the underlying human agents. This thesis is separated into three parts: (i) response dynamics in financial markets, (ii) dynamics of career trajectories, and (iii) a stochastic opinion model with quenched disorder. In Part I we quantify the response of U.S. markets to financial shocks, which perturb markets and trigger “herding behavior” among traders. We use concepts from earthquake physics to quantify the decay of volatility shocks after the “main shock.” We also find, surprisingly, that we can make quantitative statements even before the main shock. In order to analyze market behavior before as well as after “anticipated news” we use Federal Reserve interest-rate announcements, which are regular events that are also scheduled in advance. In Part II we analyze the statistical physics of career longevity. We construct a stochastic model for career progress which has two main ingredients: (a) random forward progress in the career and (b) random termination of the career. We incorporate the rich-get-richer (Matthew) effect into ingredient (a), meaning that it is easier to move forward in the career the farther along one is in the career. We verify the model predictions analyzing data on 400,000 scientific careers and 20,000 professional sports careers. Our model highlights the importance of early career development, showing that many careers are stunted by the relative disadvantage associated with inexperience. In Part III we analyze a stochastic two-state spin model which represents a system of voters embedded on a network. We investigate the role in consensus formation of “zealots”, which are agents with time-independent opinion. Our main result is the unexpected finding that it is the number and not the density of zealots which determines the steady-state opinion polarization. We compare our findings with results for United States Presidential elections.

Actions (login required)

Edit Item