The development in the working of markets has been an important topic in economic history for decades. The volatility of market prices is often used as an indicator of market efficiency in the broadest sense. Yet, the way in which volatility is estimated often makes it difficult to compare price volatility across regions or over time for two reasons. First, if prices are non-stationary, the variance is inflated. Second, the variance of commodity prices contains information on a number of region- and time-specific factors that are not related to market efficiency. Hence, the popular coefficient of variation and related indicators are not adequate measures of the efficiency of markets and are incomparable across regions. As a solution, we suggest using a conditional heteroscedasticity model to estimate the residual (conditional) variance of commodity prices. This measure reflects how markets react to unexpected events and can therefore be seen as a measure of market efficiency. Using this approach on grain prices from the Early Modern Pisa, Paris, Vienna, and Japan, we find that the residual price volatility had declined (and market efficiency increased) in the European markets in the late sixteenth century while it remained stable in Japan. © 2010 Springer-Verlag.