Threshold Cointegration and Nonlinear Causality test between Inflation Rate and Repo Rate

Katleho Makatjane,Diteboho Xaba,Ntebogang Moroke

doi:10.22610/jebs.v9i3(j).1755

Katleho Makatjane, Diteboho Xaba + Show 1 more

Open Access

https://doi.org/10.22610/jebs.v9i3(j).1755

Copy DOI

Abstract

The current study investigated a cointegration and nonlinear causality relationships between inflation and repo rates of South Africa using the data spanning the period of January 2002 to March 2016. We used a threshold vector error correction model (TVECM) and nonlinear Granger frameworks causality to carry out the analysis. Preliminary analysis of data revealed the expected properties of the data such as nonlinearity, non-stationarity and co-movement of the variables. The two variables confirmed to be moving together in the long-run according to the observed supWald test statistic. Finally, the Diks-Panchenko nonlinear Causality test revealed a strong bidirectional nonlinear causal relationship between repo rate and inflation rate. The results imply that the use of repo rate to target the inflation rate during the target period did not address the financial problem in South Africa. Consequently, the study concluded that repo rate may not be a good measure to use for controlling inflation rates of South Africa.

Highlights

There is a general impression about the presence of a linear relationship between macroeconomic variables and as such most of the empirical studies base their analysis disregarding the fact that these variables may be nonlinear (Enders and Siklos, 2001; Iyke, 2015)
The current study aims to empirically investigate the threshold cointegration and nonlinear Granger causal relationships between inflation rate and repo rate of South Africa
Threshold cointegration was estimated through the threshold vector error correction method (TVECM)

Summary

Introduction

There is a general impression about the presence of a linear relationship between macroeconomic variables and as such most of the empirical studies base their analysis disregarding the fact that these variables may be nonlinear (Enders and Siklos, 2001; Iyke, 2015). The current study proposes the application of asymmetric cointegration and nonlinear causality to inflation and repo rates of South Africa. The study contributes to academic paradigm by filling a void in practical literature by instantaneously navigating error correction modelling and causality tests focusing on an asymmetric viewpoint. Such studies are very scares in literature as far as South Africa is concerned. There is evidence of studies that used other cointegration methods on inflation rates and other macroeconomic variablesespecially linear methods but evidence of studies on this sector and repo rate with the application of nonlinear methods has not been recorded. For more readings on nonlinear causality see Nazlioglu et al (2014), Bal and Rath (2015) and Phiri (2016)

Methods

Results

Conclusion