Sectoral exposure of financial markets to oil risk factors in BRICS countries

In this paper, we examine the exposure of sectoral equity returns to changes in oil risk factors among BRICS markets. The paper combines VAR model together with the Random Forest technique to provide a novel framework which overcome some weaknesses in VAR modelling and help in the selection of oil-risk factors considered. Our results in general show that the response of sectoral stocks to changes in oil-risk factors is not homogenous and their sensitivities are not uniform across BRICS markets. Stock returns of Basic Materials, Financials and Industrials sectors show consistently negative sensitivity to crude oil volatility (OVX) and oil price shocks for major oil importers, whereas sectoral stocks among major oil exporters seems to demonstrate systematic pattern. To explain sectoral stock variations, the results reveal OVX and oil to be the most significant risk transmitting factors in BRICS, irrespective of whether the market is a major oil importer or exporter. However, our results also confirm the presence of asymmetric effects in sectoral stocks for Brazil and Russia while no such evidence is found for India, China and South Africa. Overall, the findings suggest that the manner in which sectors respond to fluctuations in oil-risk factors varies between different sectors and it depends on the oil characteristics of the markets, nature of sectors and the type of oil-risk factors considered. •We examine the exposure of sectoral stock returns to oil risk factors among BRICS markets.•We introduce oil volatility index to account for oil market uncertainty and Random Forest model for importance of variables.•We find that the sensitivity of sectoral stocks to oil-risk factors is not uniform with heterogenous exposure across markets.•Oil volatility index and oil price are found to be the most significant risk transmitting factors to in BRICS countries•Oil-risk factors show asymmetric effect on sectoral returns for major oil exporting markets (Russia and Brazil).•Random Forest-VAR model provides methodological contribution to the literature.