MIRCOECONOMETRICS: TIME SERIES ANALYSIS AND MODEL DEVELOPMENT OF RELATIVE INFLATION AND CONSUMPTION TRENDS
Open Access
Author:
Kapil, Alvin Milan
Area of Honors:
Statistics
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
David Russell Hunter, Thesis Supervisor David Russell Hunter, Thesis Honors Advisor Dr. Zhibiao Zhao, Faculty Reader
Keywords:
Statistics Inflation Econometric.
Abstract:
In this observational study, different subspaces of inflation based on geographical and class factors were compared for significance and modeled through the following time series analysis (TS) methods: the autoregressive integrated moving average (ARIMA) model and the generalized autoregressive conditional heteroskedasicity model (GARCH). The geographical factors were defined by four separate regions (Northeast, West, Midwest, and South) and were tested for significance by an analysis of variance (ANOVA) on the major individual components of an American’s budget: housing, energy, food, medical costs, and apparel. The class factors were also tested for significance by ANOVA based on the income for the following three groups defined by an academic class model: upper-middle class, lower middle class, and the working class. Next, each component was explored by both ARIMA and GARCH, either individually by class or geographical factor if the ANOVA was significant, or nationally if the ANOVA was not. By optimizing each method for a given parameter, applicable matrices of inflation rates were generated to understand the phenomena for relative inflation. The implications for accurate relative inflation could provide Americans a better understanding of their personal finances and an additional tool for expense planning.
