Publication: Forecasting Consumer Price Index in the United States
Type:
Article
Date
2023-11-01
Journal Title
Journal ISSN
Volume Title
Publisher
Faculty of Humanities and Sciences, SLIIT
Abstract
This report presents the Auto-Regressive
Integrated Moving Average (ARIMA) model for
forecasting the consumer price index (CPI) in US
using monthly data from March 2010 to March
2023. The original series was not stationary,
but the first difference series was found to be
stationary using the Augmented Dicky Fuller test.
The best-fitted model was identified based on the
significance of the parameters, volatility (sigma2),
log-likelihood, Akaike, Schwartz, and Hannan-
Quinn information criterion. Parameters of the
fitted model are significantly deviated from zero.
The stability of the model has been checked using
the roots of the unit root test. Residuals of the
fitted model satisfied the randomness but nonconstant
variance. The monthly forecasted values
of CPI from April 2023 to August 2023 are 301.833,
302.444, 303.038, 303.639, and 304.261. The
percentage errors of the forecasted values are less
than one percent. This method and results provide
useful information to policy and market makers
for their planning,
Description
Keywords
ARIMA, Consumer price index, Forecasting, Box Jenkins, Time series analysis
Citation
Witharana, W. W. S. K., Udugama, U. K. D. T. N., Fernando, P. M. R., Kaumadi, H. M. H., Peiris, T. S. G. (2023). Forecasting Consumer Price Index in the United States. Proceedings of SLIIT International Conference on Advancements in Sciences and Humanities, 1-2 December, Colombo, pages 290- 295.