Title: Forecasting inflation in Tunisia during instability using dynamic factors model: a two-step based procedure based on Kalman filter

Authors: Bilel Ammouri; Hassen Toumi; Fakhri Issaoui; Habib Zitouna

Addresses: University of Tunis (UR-DEFI and ESSEC Tunis), 4, Abou Zakaria El Hafsi Street, Montfleury, Tunis, 1089, Tunisia ' College of Business Administration, Taibah University, Yanbu, Kingdom of Saudi Arabia; University of Economics and Management of Sfax, (URED), Street of Airport km 4.5, LP 1088, Sfax 3018, Tunisia ' University of Tunis-El Manar (LR-PS2D) and University of Tunis (ESSEC Tunis), 4, Abou Zakaria El Hafsi Street, Montfleury, Tunis, 1089, Tunisia ' University of Carthage, UR-MASE and Faculty of Economic Sciences and Management of Nabeul, Campus Universitaire Mrezga, Nabeul, 8000, Tunisia

Abstract: This work presents a forecasting inflation model using a monthly database. The model has to take into account a large amount of information, is the goal of recent research in various industrialised countries as well as developing ones. With the dynamic factors model (DFM), the forecast values are closer to the actual inflation than those obtained from the conventional models in the short term. In our research, we devise the inflation into 'free and administered' and test the performance of the DFM under instability in different types of inflation (core and trend). Knowing that periods of instability are simultaneously the period of price liberalisation of basic goods (2008) and the post-revolution (the Arabic spring) period (2011-2014). We have found that the DFM with an instability factor leads to substantial forecasting improvements over the DFM without an instability factor in the period after the revolution.

Keywords: inflation forecasting; PCA; principal component analysis; VAR; vector autoregressive; DFM; dynamic factors model; Kalman filter; space-state; instability factor.

DOI: 10.1504/IJCEE.2019.097794

International Journal of Computational Economics and Econometrics, 2019 Vol.9 No.1/2, pp.49 - 83

Received: 28 Nov 2016
Accepted: 26 Jun 2017
Published online: 11 Feb 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article