The Granger causality between confidence and consumption in South Africa
Date
2018
Authors
Singh, Vimal
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
South Africa’s news headlines are dominated by controversial stories of corruption, crime and politics. This research report investigates if people always accept these events as a "normal” part of the country’s history, or if these factors influence or are influenced by expenditure decisions of businesses and consumers. The variables included in the investigation are household consumption, business capital formation, consumer confidence and business confidence. The investigation establishes that these variables are non-stationary and cointegrated, with the cointegrating relationship assessed using Johansen’s procedure. The short-run and long run dynamics between the variables are determined using vector error correction models. Granger causality tests were used to explore the causal relationship between the variables.
The Granger Causal relationship between confidence and consumption is assessed using quarterly data from June 1982 to March 2017. It showed that changes in household consumption Granger cause changes in consumer consumption, and no such relationship exists between business confidence and capital formation. The Granger Causal relationship between confidence indicators was also explored, which found that a bi-directional Granger causality relationship existed between business confidence and consumer confidence.
The results of variance decomposition (VDC) and impulse response functions (IRFs) were applied thereafter to further examine the causal relationship between the variables. The former determines the amount each variable contributes to each other while latter assess the impact on the dependent variable given a shock to the system. The results supported the outcome of the Granger causality tests. The variance decomposition found in most cases that a shock to the dependent variable can explain more of the forecast error in the dependent variable than a shock to the other predictor variable. This was observed in the short and long run. The impulse response functions found that confidence measures, both for consumers and businesses, may respond in the initial periods to impulses but the increments of the increase reduce after 1 to 2 periods.
Description
A research report submitted to the Faculty of Science, in partial fulfilment of the requirements for the degree of Master of Science, University of the Witwatersrand. Johannesburg, April 2018.
Keywords
Citation
Singh, Vimal (2018) The Granger causality between confidence and consumption in South Africa, University of the Witwatersrand, Johannesburg, https://hdl.handle.net/10539/26058