Financial Innovation

Frame W S and White L J (2002) “Empirical Studies of Financial Innovation: Lots of talk, little action “WP 2002-12 Federal Reserve bank of Atlanta
DIFFUSION
We are aware of five studies of the diffusion of financial innovations, three of which
focus on ATM deployment by banks. These studies generally use hazard models that estimate
the adoption pattern of the innovation under study conditional on firm- and market-specific
effects.
Hannan and McDowell (1984) find that -- consistent with the Schumpeterian hypotheses
-- larger banks and those operating in more concentrated local banking markets registered a
higher conditional probability of ATM adoption. This study also found bank product mix, bank
holding company affiliation, urban location, branch banking restrictions, and the area wage rate
were all positively related to ATM adoption.
In a subsequent study, Hannan and McDowell (1987) find that the conditional
probability of ATM adoption is positively related to a rival's adoption and that firms in less
concentrated markets react more strongly to rival precedence than do their counterparts in
concentrated markets. Consistent with their previous results, bank size and local market
concentration were positively related to ATM adoption. Similar results were found for bank
19
holding company affiliation, branch banking restrictions, and market deposit growth.20
Using the same data, Saloner and Shepherd (1995) find that the expected time to
adoption of ATMs declines in both the number of users (deposits) and locations (branches),
indicating the presence of network externalities. For limited branching states, market
concentration is positively related to ATM adoption speed, while depositor growth is negatively
related. For unrestricted states, the area bank wage rate is positively related to ATM adoption
speed.
Molyneux and Shamroukh (1996) examine the diffusion of the underwriting of junk
bonds and of note issuance facilities (NIFs) during the...