Analysis of US Per Capita Income Growth Using Markov Chains

Open Access
Mcgarry, Brynne Marie
Area of Honors:
Interdisciplinary in Economics and Statistics
Bachelor of Science
Document Type:
Thesis Supervisors:
  • David Russell Hunter, Thesis Supervisor
  • John Fricks, Honors Advisor
  • Russell Paul Chuderewicz, Faculty Reader
  • Markov chain
  • economic growth
  • statistics
  • change points
  • maximum likelihood estimation
Despite its constant presence in the news, the discussion of growth theory in economics rarely emphasizes its deep mathematical aspects. While our nation’s economic history combined with recent social and political events provide the driving forces behind this growth, the story can be verified quantitatively. This paper considers one method for looking at economic growth in the US through a technical lens; namely, we treat changes in income growth over the last 82 years as a Markov chain and use maximum likelihood estimation to find time points where changes appear most drastic. After exploring the data and discussing some of the finer points of Markov processes, the algorithm for finding one change point is discussed in depth and extended to two change points and the general case.