This dissertation consists of three essays. In the first essay, econometric models are used to measure price risk in a study for major grains (wheat, rice, corn, and soybeans) in China. Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models and Multiplicative Heteroskedasticity (M Het) models are applied to estimate time-varying price variance, and then covariances are estimated by a simple two-step process assuming constant conditional correlations. An aggregate price risk index is constructed from these variances and covariances using an economic index number approach. In theory, this approach is superior to the more common approach of estimating a univariate GARCH model for an aggregate price index. This easay compares the two approaches to measuring aggregate price risk and finds low correlations. Thus there is substantial difference between the two approaches in practice as well as in theory. The previous essay measures aggregate price risk but does not explain price risk. The second essay attempts to investigate potential factors that contribute to aggregate price risk of major grain products (rice, wheat, corn and soybeans) on monthly base in China from mid 1980s to recent year from both theoretical and empirical perspectives. The superlative price risk indexes are explained by a set of key variables that characterize China’s economy, agricultural market and trade as well as biological system of major grain in China. These variables account for much of the variation in the aggregate price risk index. Moreover empirical results favor use of the superlative index of aggregate risk rather than standard measures of aggregate risk. The third essay is an extension of previous two essays by explaining price risk at disaggregate level. Price variances and covariances are modeled using both Ordinary Least Squares (OLS) and Seemly Unrelated Regression (SUR) techniques. Results are broadly consistent with the previous essays.