Economics Seminar - "Missing Financial Network Data"

Little data exists describing the links of the US financial network. Using a computational model of interbank lending, I show that this lack of data can lead to erroneous model predictions. I analyze several types of data inaccuracies and find that missing a single loan in the network can lead to large differences in the predicted number of unpaid loans and total dollars repaid. This missing data could mean implementing policies that are designed to improve macroeconomic stability but that could actually lead to substantial destabilization. These results are robust across multiple network sampling regimes.