Chad Fulton

Education and Research

I am a fifth year PhD student in the economics department at the University of Oregon. My research focuses on rational inattention, considering boundedly rational agents who cannot perfectly track economic conditions and must choose what to pay attention to. I apply these theoretical tools along with large-dimensional econometric tools (for example dynamic factor models) to explain the behavior of price movements at the firm, sector, and aggregate levels.

Experience

Google Summer(s) of Code: Time-Series Econometrics in Python

I spent three summers (2013, 2014, and 2015) participating in Google Summer of Code on behalf of Statsmodels, an open source project bringing statistical and econometric modeling tools to the Python programming language. For more information, see the Code section. In brief, my contributions included:

• (2013) Non-linear time series estimation: Developed software to estimate TAR, STAR, and Markov switching models, along with statistical tests for non-linearities and for breakpoints with unknown timing.
• (2014) State space models: Developed software to specify and estimate arbitrary state space models using an efficient (compiled) Kalman filter with convenient Python wrapping classes for syntatic sugar. As an application, developed specific classes allowing the estimation of SARIMAX models.
• (2015) State space extensions: Improved the state space estimation routines by improving standard error calculation, adding diagnostic tests and plots, adding impulse response functions and time series simulation, adding the Kalman smoother, and adding unobserved components, VARMAX, and dynamic factor models.

---

ECONorthwest and Evergreen Economics:

Between undergraduate and graduate school, I spent two years "getting my hands dirty" performing economic analysis for ECONorthwest and Evergreen Economics, two economic consulting firms operating out of Portland, Oregon. I conducted or supported a number of interesting projects, including:

• Connected lighting: Developed software to robustly collect, manage, and analyze large volumes (tens of millions of records) of high-frequency (15-second) interval data on a real-time basis from Philips Hue devices installed in residential homes.
• Impact analyses: Performed econometric modeling of energy savings associated with equipment provided to eligible low-income households. Constructed, cleaned, and analyzed the large-scale relational datasets using SAS.
• Conjoint analysis: Developed a sample and web-based survey instrument and performed econometric modeling for a conjoint analysis of utility customers' preferences related to energy efficiency measures.