ESROP – Permanence and Periodicity: An Analysis of Structural and Cyclical Changes in Macroeconomic Aggregates

With the rise of protectionist policies in the United States, rapid integration of artificial intelligence in business, and demographic pressures from an aging population, Canada finds itself at the intersection of several powerful economic forces. These dynamics blur the distinction between long-run structural change and short-run cyclical fluctuations in macroeconomic data, increasing the risk of misinterpretation and policy miscalibration.

This project seeks to develop modernized analytical models that more effectively disentangle structural and cyclical components in macroeconomic data by leveraging time-series unsupervised learning models and graph neural networks. Existing economic models rely heavily on traditional statistical decomposition methods and the adoption of machine learning remains limited and transitional. This research addresses that gap by integrating advanced time-series machine learning with established macroeconomic theory.

Building on the combined expertise of researchers at KMUTT and the University of Toronto, the project employs a multi-phase methodology that pairs machine-learning model development with social-science case studies. The result is an advanced model designed to improve macroeconomic analysis while fostering research collaboration between Canada and Thailand.

Ultimately, this research aims to strengthen the analytical tools available to central banks, supporting sound monetary policy decisions, economic resilience, and confident navigation of an increasingly complex global economy.

Faculty Supervisor:

Arthur Chan

Student:

Partner:

King Mongkut’s University of Technology Thonburi

Discipline:

Sociology

Sector:

Public Service, Policy, and Governance; Finance and Insurance

University:

University of Toronto

Program: