Commodity and energy prices have exhibited an unprecedented increase between October 2006 and July 2008, only to fall sharply during the last months of 2008. Many explanations have been offered to this phenomenon, including steadily increasing demand from China and India, large mandated increases in ethanol production, droughts in some key agricultural producer countries, production plateaus in some major oil-producing countries, refinery capacity limits, demand pressure from the derivatives market owing to the diversification properties of commodities, etc. Clearly, agricultural input, output, and energy products are closely related economically. In addition to biofuels, the connection points include nitrogen-based solution liquid fertilizers, fossil fuels used in agricultural production, limited acreage available for field crops, etc. While all these price variables are, evidently, closely connected, it is not entirely clear how exogenous price shocks are transmitted through the system, and whether particular commodities drive up the prices of other commodities. The proposed paper attempts to address this "chicken-or-egg" problem by applying the Structural Vector Autoregression (SVAR) framework to the analysis of a wide range of commodity prices. The purpose of this paper is to model and estimate an SVAR model of multiple commodity prices in order to: (i) Investigate the transmission mechanism of the price shocks associated with the commodity boom of 2006-2008 and subsequent bust and identify changes (if any) relative to earlier time periods (e.g. 2004-2006), (ii) Evaluate the possibly asymmetrical relationship between different commodity and energy price variables and (iii) Test hypotheses of causality in a time series definition (Granger and graph-theoretic). The methodology consists of defining and estimating a structural VAR model, studying impulse response functions and the variance decomposition, and testing for the direction of causality. Given that a longstanding problem is the sensitivity of the results to identifying assumptions, graph analysis is used here as it is a promising approach to identify the structure of the variance-covariance matrix and therefore overcome the observational equivalence of various reduced-form models implied by different identification approaches. The main contribution of this paper is to provide a tentative answer to the question of how agricultural input, output and energy product prices are related and whether this relationship has changed in recent years.