In the real option pricing model of valuation and decision making, the estimation of future volatility is a key input parameter. For traded commodities or financial assets, past volatility is used as a proxy for predictions. But, for projects, this approach is not feasible because, in most cases, historical data of traded projects are not available. As an alternate solution, it is usually assumed that project volatility is equal to that of commodity price. In order to investigate this assumption, we estimate the project volatility considering that both commodity price and operating cost evolve as a geometric Brownian motion (GBM). Results of a hypothetical gold mining project indicate that project volatility is higher than that of commodity price and it only drops to price volatility under very unrealistic industry conditions, such as very high prices or very low production costs. in addition. we find that project volatility is independent of production capacity and taxation, but depends on increments in price and cost, as well as strongly on their degree of correlation. (c) 2006 Elsevier Ltd. All rights reserved.3128694