Spillovers from entry: the impact of bank branch network expansion

• DOI: http://doi.org/10.1111/1756-2171.12258
• Date: 01 December 2018
• Published date: 01 December 2018
• Author: Joseph Kuehn

RAND Journal of Economics

Vol.49, No. 4, Winter 2018

pp. 964–994

Spillovers from entry: the impact of bank

branch network expansion

Joseph Kuehn∗

I study why local banking marketsbecame dominated by multimarket firms following deregulation

in the 1990s. I estimate a model of branch entry that allows for spillovers across markets. The

spillovers complicate estimation, and so I develop a revealedpreference approach that also deals

with unobserved firm and market heterogeneity. I then analyze the impact of multimarket banks

and find that they increase local competition, but that they also open more branches than single-

market firms, and subsequently offer lower deposit rates. Ultimately, their welfare impact differs

across markets based on the availabilityof outside alternatives and consumer sensitivity to rates.

1. Introduction

In many industries, local market competition has become increasingly dominated by firms

that operate in multiple geographic markets. Examples include retail, insurance, telecommuni-

cations, and banking. In this article, I study how the advantages attained by multimarket firms

(e.g., economics of scale, vertical differentiation) affect competition in local markets that are

increasingly comprising these firms, specifically their effect on firm size, pricing, and ultimately,

welfare.

In particular, I look at the banking industry. Historically, banks were restricted from expand-

ing their branch networks beyond a local geographic area. Throughout the 1980s, many of these

regulations were removed. Since that time, banks have gradually expanded into new markets,

by both adding new branches (the number of bank branches has increased from around 50,000

branches in 1990, to over 80,000 branches in 2010), and through consolidation (the number of

banking institutions has declined from around 12,000 in 1990, to 6,500 institutions in 2010).

The profusion of multimarket banks suggests that there are important economic benefits to

operating in multiple markets. These may include larger branch networks attracting more con-

sumer deposits, reduced costs due to economies of scale, and a diversification of risks geographi-

cally.My first main objective is to understand what has driven the postderegulation expansion into

multiple markets. I do this by identifying the impact each of these potential spilloversources have

∗CSU East Bay; joseph.kuehn@csueastbay.edu.

I am grateful to the Editor, AvivNevo, and two anonymous referees for their valuablecomments and suggestions. I would

like to thank Connan Snider for his extensiveadvice and suppor t. I wouldalso like to thank John Asker and Jinyong Hahn

for their guidance, and Hugo Hopenhayn, John Riley,Rosa Matzkin, Dan Bernhardt, Siwei Kwok, and Megan Accordino

for their helpful comments and suggestions. All errors are my own.

964 C2018, The RAND Corporation.

KUEHN / 965

on local bank profitability. The second objective is to then quantify whether markets comprising

firms operating on a national level are substantively different in terms of competition, market

outcomes, and welfare, than if the local firms had no presence in outside markets. I am particularly

interested in how national banks have affected the total number of banks a market can sustain,

their sizes in terms of branches, the deposit rates they set, and their effect on consumer and firm

welfare. In answering these questions, this article also contributes to developing a technique to

better understand the importance of spillovers in entry models, and how firms make decisions

over an entire network of markets.

To understand what has driven branch expansion, I first look at how a large branch network

affects a bank’s ability to attract consumer deposits. I do this by estimating a model of consumers’

choice over which bank to place their deposits in. This choice depends on the deposit rate offered

by the bank, the consumer’s dollar value of deposits, and the bank’s networkof branches, including

the number of branches and their locations. Estimating the model using the approach of Berry,

Levinsohn, and Pakes (1995), I find that the median consumer wouldbe willing to accept a nearly

33% drop in the average deposit rate in exchange for one additional branch within their market.

On the other hand, they care very little about a bank’s number of branches outside of their home

market.

To quantity additional incentives for branch expansion, such as economies of scale and risk

diversification, I estimate a model of bank branch entry. Branch entry is modelled as each bank

chooses the number of branches to open in each county in the United States to maximize their

total profitability across all markets. Bank profits depend on the variable profits they receivefrom

collecting deposits at the branches, which are calculated using the estimates from the demand

model, and the costs of adding the branches. I specify costs as a function of the number and

locations of the bank’s existing branches outside the market.

With cross-market spillovers, a bank’s actions in one market have ancillary effects on their

profitability of opening branches in other markets. This complicates finding an equilibrium

solution, as each bank’s decision on the number of branches to open in each market is related to

the decisions of competing firms in that market, and their decisions in over 3000 other markets.

To deal with this complexity, I use a revealed preference approach that builds on the moment

inequality technique of Pakes, Porter, Ho, and Ishii (2015). I use data on the location of each

branch in a bank’s network to infer what parameter vector is consistent with the observed branch

choices being profit maximizing. I control for both unobserved market-specific and firm-specific

effects, so that such heterogeneity is not mistakenly attributed to positive branching spillovers.

I find that there is a substantial cost advantage to operating branches in multiple markets.

Having 10 additional branches in outside markets within the state increases the profitability of

adding a branch by somewhere between 3.5% and 4.2% of average per-branch variable profits. I

also find that this advantage is larger for more expansive networks. This suggests that it mainly

comes from diversifying geographically.

To investigate the impact of multimarket banks, I run counterfactual experiments that elim-

inate the estimated spillovers. The first counterfactual removes all cross-market spillovers. The

results show that a little over30% of the branches b uiltin the obser ved outcome can be explained

by these cross-market incentives. Expansion leads to a decrease in local market concentration,

but average deposit rates actually go down. This is because multimarket firms have lower costs

to building branches, which increase their market power, and allow them to lower deposit rates

despite facing more competitors. Expansion is welfare increasing in rural and small city mar-

kets, and welfare decreasing in more populous markets, which differ for two main reasons. First,

competition from the outside option is stronger in larger markets due to the greater availability

of alternatives to banking in these markets. Second, larger markets have a higher proportion of

consumers with large deposit endowments, whom are more sensitive to changes in the deposit

rate, and care less about the size of a bank’s branch network, compared to the average customer.

As expansion pushes out smaller banks that generally offer better deposit rates, the rich customers

in these markets tend to then switch not to the remaining multimarket banks, but instead to the

C

The RAND Corporation 2018.