The impact of oil shocks on the South African economy

ABSTRACT

The recent increases in oil prices have raised the importance of studying the effects of oil supply and demand shocks on an economy. The purpose of this paper is to investigate the impact of the oil supply and demand shocks on the South African economy using a sign restriction-based structural Vector Autoregressive (VAR) model. The results of this study show that an oil supply shock has a short-lived significant impact only on the inflation rate, while the impact on the other variables is statistically insignificant. Supply disruptions result in a short-term increase in the domestic inflation rate with no reaction from the monetary policy. An aggregate demand shock results in short- to medium-term improvements in domestic output and the real exchange rate. The effect is statistically insignificant for the inflation rate as well as the monetary policy instrument. The inflation rate and the real exchange rate react negatively to an oil-specific demand shock, while output is positively related to unanticipated changes in oil price due to speculations. This study’s results highlight the importance of understanding the source of the oil price movements, since an oil price increase necessarily does not imply a negative effect on the economy.

KEYWORDS:

• Macroeconomic variables
• monetary policy
• oil price shocks
• vector autoregression

Acknowledgments

The authors would like to thank Professor Gert Peersman, the Editor-in-Chief, Dr. James G. Speight, and an anonymous referee for many helpful comments. Any remaining errors are, however, solely ours. A longer version of the paper, which includes unit root tests and international comparison of our results, is available at: http://web.up.ac.za/sitefiles/file/40/677/WP_2013_11.pdf.

Notes

¹ For further details on the construction of this index, the reader is referred to Kilian (2009).

² The details of the unit root and cointegration tests are available upon request from the authors.

³ For a detailed discussion of the methodology, the reader is referred to Peersman and van Robays (2009).

⁴ We also estimated IRFs based on a VAR with one lag (suggested by the Schwarz Information Criterion and the Hannan–Quinn Information Criterion) and four lags (selected by the sequential modified likelihood ratio test statistic) as well, and found the results to be relatively similar to the ones with two lags reported in the paper. The results based on one lag and four lags are available upon request from the authors.

⁵ We carried out two subsample analyses starting 1990:Q1 and 1994:Q1 to capture, respectively, the beginning of informal inflation targeting by the SARB and South Africa’s integration in the world economy, which, in turn, led to a sharp increase in oil consumption, to check for the robustness of our results in terms of the interest rate behavior. Our results, however, continued to indicate lack of significant response of the SARB following the different shocks that led to an increase in oil price. The details of these results are available upon request from the authors.