Using Directed Acyclic Graphs (DAG's) and Error Correction Models we study the dynamics of the notoriously volatile international freight prices that comprise the Baltic Panamax Index, the index on which freight futures trading is based. The DAG's are used to make definitive statements about the contemporaneous correlations between prices and allow us to address the construction of the data-determined orthoganization on contemporaneous innovation covariance, critical in providing sound inference in innovation accounting techniques. Our results provide a rich source of information on price discovery over various time horizons and suggest that the index may not be appropriately comprised and weighted.