Network DEA provides a flexible way to customize DEA problems to specific applications. It provides 'links' across DMU's, or alternatively allows us to look inside a complex DMU with multiple nodes. In this short paper, we provide an illustration of a dynamic network DEA model. It is a network in the sense that it links the behavior of DMU's across time. Our particular illustration is to an 'old' issue: the impact of public capital on technology and productivity (TFP) at the state level in the U.S. We use a dynamic network production model developed by Fare and Grosskopf [8,9] to measure dynamic efficiency. As a byproduct of the model, we solve for optimal public and private investment paths, which we then can compare to realized investment. We apply this to U.S. manufacturing data over the 1978-1999 period using state level data.
