Separation of precious metal beta from a JSE multivariate model with macroeconomic variables

• Main Author: Mzobe, Thabani Bonginkosi
• Other Authors: Holman, Glen
• Format: Dissertation
• Language: English
• Published: University of Cape Town 2016
• Subjects: Financial and Risk Management
• Online Access: http://hdl.handle.net/11427/18205

Includes bibliographical references === This study examines a multifactor model of the Johannesburg Stock Exchange (JSE) framed within the Arbitrage Pricing Theory (APT). The APT has been set up such that it can be able to separate the beta for the precious metal factor within the model. The process goes via the investigation of macrovariables (with precious metals used as one of the macrovariables) and their effect on market (JSE) returns. A complete analysis and modeling of this relationship is likely to yield unparalleled rewards and cost-effective risk management, monitoring and mitigation. Using monthly data for the period 31/07/2002 to 30/04/2013 the dissertation focuse d on using a market (JSE) representative index as a basis for creating a wholly functioning APT model. This included creating a more liquid representative of the JSE All Share Index (A LSI) by using the top 100 stocks by market capitalization. Principal Components Analysis (PCA) was applied to the variables to ascertain a proper model for the JSE return structure. However, in the end an appropriate econometric structure in the form of Autoregressive Conditional Heteroscedastic (ARCH) and Generalized Autoregressive Conditional Heteroscedastic (GARCH) models was used and applied to test and create the APT model to address the objective. The other purpose of this dissertation was to separate beta attributable to the precious metal macrovariable within the model. This is based on the establishment of the JSE in the late 1880s being primarily due to the discovery of precious metals in the former Transvaal (North West) and Pretoria, Witwatersrand and Vereeniging (PWV) region now Gauteng. This is to ascertain whether these metals still have as much influence on the JSE as they did for over half a century. The results show that macroeconomic variables do influence the return generating process of the JSE, explaining almost 80% of variation in returns. The results show that the ALSI is characterized by a seven factor APT with, industrial production, money supply, SA consumer price index, ZARUSD exchange rate, crude oil, MSCI ACWI and precious metals statistically significant.