Review of Industrial Organization 20: 349–365, 2002.
© 2002 Kluwer Academic Publishers. Printed in the Netherlands.
Predicting Household Switching Behavior and
Switching Costs at Depository Institutions
ELIZABETH K. KISER
Federal Reserve Board, 20th and C St., NW, Washington, DC 20551, U.S.A.
Abstract. This paper uses new survey data to investigate the covariates of self-reported switching
costs and switching behavior by deposit account holders. Factors affecting geographic mobility ap-
pear to be most important in explaining the duration of deposit relationships. Both younger and older
respondents are more likely than others to be at their ﬁrst bank ever, suggesting a cohort effect in
deposit relationships. Households reporting switching costs, net of the beneﬁts from switching, are
less likely than others to have stayed with a bank for prices or customer service, suggesting that
switching costs may decrease price sensitivity. Switching costs appear more severe for households
with high income or education and for households with very low income or minority ethnicity. These
ﬁndings imply that banking markets characterized by such households may present greater entry
costs for new ﬁrms.
Key words: Depository institutions, merger policy, switching costs.
Customer substitution across ﬁrms and products is central to merger analysis.
Estimates of cross-price elasticities of demand are commonly used to deﬁne geo-
graphic and product markets. In addition to questions of market deﬁnition, merger
policy in retail banking often relies on arguments stating the degree of potential
competition in a market – that is, whether ﬁrm entry would likely occur if existing
ﬁrms were to raise prices or reduce quality. Because a ﬁrm’s success of entry hinges
on its ability to attract new customers, customer switching is extremely important
for the viability of new entrants.
This paper uses new household survey data to study the switching behavior of
holders of deposit accounts. The survey is unique in that it asks households detailed
questions about their relationship to their depository institution, including their
reasons for remaining with an institution. Regressions are performed to predict (1)
household tenure at the deposit bank, (2) whether the household has ever changed
The views expressed in this paper are those of the author and not necessarily those of the Federal
Reserve Board. The author would like to thank Bob Adams, Dean Amel, Marianne Bitler, Myron
Kwast, Robin Prager, Steve Rhoades, Paul Smith and two anonymous referees for helpful comments.