Mardi | 2020-06-02
11H00 – Webinar
AbstractWe construct an unique microgeographic data to study the effects of Canadian bank branch network on consumer’ cash management. We create the average distance of consumer’s travelling to (affiliated) closest branch on the Forward Sortation Area (FSA) level, which avoids the attenutation bias and accounts for spatial bank branch network. By shrinking the difference of distances between general Financial Institution (FI) and specific FI branch, we are able to separate the effects of shoe-leather cost from withdrawal fee. We find the significantly threshold effect of consumer-branch distance on the typical month withdrawal frequency: the first kilometer plays the most important role, the coefficient around negative 50 percent. Moreover, the estimates are consistent with various model specifications and robust with respect to (a) consumer residential sorting; (b) indirect network effect of withdrawals on bank branch network; and (c) misclassification of branch withdrawals driven by the automated teller machine (ATM) network.