Time series techniques—particularly impulse–response functions and variance decompositions—are used to characterize the short-run relationships between 17 variables in a vector autoregressive model designed to trace the short-run interconnections among variables affecting lockages on the Mississippi and Illinois Rivers. The model contains five categories of variables: lockages, barge rates, grain bids, rail rates, and rail deliveries. Variance decompositions are constructed that identify barge rates as the most important variable affecting lockages at both short and long horizons. Barge rates are, in turn, explained largely by lockages and rail rates, indicating two-way feedback or bidirectional causality between lockages and barge rates. Impulse–response functions are also examined. The variance decompositions indicate that barge rates are important in explaining lockages, and the impulse–response functions show how lockages and other variables respond to such shocks. In general, there is a substitution away from barge transportation and toward rail transportation when barge rates increase. The results are useful for illuminating the causal relationships among variables in the model and for understanding behavioral relationships present in the data and can be used to guide short- and long-run planning models. For example, many planning models assume that barge traffic does not respond significantly to changes in barge rates; however, results obtained here imply that barge traffic and rail deliveries do respond to such changes. This potentially important implication illustrates the usefulness of the time series techniques used.
Impact of credit risk and profitability on liquidity shocks of Namibian banks: an application of the structural VAR model