Measuring risk in the mining sector with ARCH models with important observations on sample size

This paper uses the GARCH technique to estimate time-varying individual firm risk measures for the mining sector. In general, the mining industry is riskier than the market with estimated betas greater than one. The results also show that the level of risk in the industry is quite volatile and can be divided into three distinct time periods according to the magnitude and variability of the firm's betas. Additionally, the number of observations in a sample has a strong relationship to the detection and estimation of an ARCH effect in the data, which cannot be explained by firm financial size.