Price risk is estimated for a representative UK arable farm using value-at-risk (VaR). To determine the distribution of commodity returns, two multivariate generalised autoregressive conditional heteroscedasticity (GARCH) models, with t-distributed and normally distributed errors, and a RiskMetrics (TM) model are estimated. Returns show excess kurtosis and that the GARCH model with t-distributed errors fits best. Estimates of VaR differ between models: both GARCH models perform well but the RiskMetrics (TM) model underestimates expected losses. UK arable farms face substantial price risk.