DOES BANK TECHNOLOGY AFFECT SMALL BUSINESS LENDING DECISIONS?

• Published date: 01 March 2017
• Author: John Sedunov
• Date: 01 March 2017
• DOI: http://doi.org/10.1111/jfir.12116

DOES BANK TECHNOLOGY AFFECT SMALL BUSINESS

LENDING DECISIONS?

John Sedunov

Villanova University

Abstract

I examine the effect that technology has on soft-information lending and address issues

within the banking literature on quantifying bank technology. I find that banks engage in

less soft-information lending when back-office bank technology is more productive and

that banks engage in less soft-information lending when they own interactive web

technology. I find that competition, lending decisions, and bank size are the primary

drivers of technological development. I show that these results are robust to econometric

tests that account for endogeneity, to an alternative definition of the bank’s size, and to

the inclusion of lending portfolio controls.

JEL Classification: G20, G21

I. Introduction

I study the impact of bank technology on the availability of credit to soft-information

borrowers in the United States. From 2001 to 2008, banks devoted an increasingly

smaller portion of their lending portfolios to soft-information borrowers. According to

Petersen (2004), soft information includes “opinions, ideas, rumors, economic

projections, statement of management’s future plans, and market commentary”(p. 6).

Moreover, this type of information is generally not quantifiable. One example of the

unquantifiable nature of soft information is the character of the borrower. Character is

part of the knowledge base that a local relationship lender might have and use.

Accordingly, I use small business lending as a proxy for soft-information lending, as

small businesses must often rely on unverifiable information to secure funding (Stein

2002). Empirically, I find that banks with greater technological capabilities lend less to

small businesses. I refine this question to separate the effects of large banks ($1 billion or

more in assets) and small banks. I find that larger banks also decrease their small business

lending when they possess better technology but in a lower magnitude. These findings are

robust to simultaneous regression models that take into account the fact that large banks

may consciously avoid small business lending because of its higher cost structure and the

economies of scale that large banks enjoy in making loans to larger borrowers.

I thank Allen Berger, Lamont Black, Gary Bliss, Ken Cyree (associate editor), Isil Erel, Scott Frame,

Bernadette Minton, Mike Pagano, Mitchell Petersen, Yingying Shao, Ren

e Stulz, and participants at the 2014

Eastern Finance Association annual meetings and 2014 Southern Finance Association annual meetings for helpful

comments and feedback.

The Journal of Financial Research Vol. XL, No. 1 Pages 5–32 Spring 2017

5

© 2017 The Southern Finance Association and the Southwestern Finance Association

RAWLS COLLEGE OF BUSINESS, TEXAS TECH UNIVERSITY

PUBLISHED FOR THE SOUTHERN AND SOUTHWESTERN

FINANCE ASSOCIATIONS BY WILEY-BLACKWELL PUBLISHING

Technological capability is composed of two parts: front-office technology and

back-office technology. Front-office technology, such as an interactive website, allows

customers to conduct banking business without having to enter the bank. Back-office

technology is a measure of how efficiently information can flow through the bank.

Following Petersen and Rajan (2002), I measure back-office technology by standardizing

the total employees of the bank by total loans.

1

I include both forms of technology in my

tests and show that both front-office and back-office technology have a similar effect on

small business lending. Namely, I estimate that a one-standard-deviation increase in a

bank’s technological productivity is related to a decrease in small-business lending of

3.36% and that the presence of an interactive website is related to decreases in small

business lending of 3.50%.

My research questions are pri marily motivated by three stud ies. First,

Williamson (1988) describes organizational diseconomies that emerge within an

organization operating in mul tiple business lines. This m ay be true for large banks,

which as a result may shift exclusi vely to hard-information lendi ng and away from

soft-information lending to avoid such diseconomies. Second, Berger and Udell

(2002) show that soft-infor mation lending is difficult with multiple layers of

management. Large banks, which may be mor e technologically capable, oft en have

multiple layers of manage ment. Third, Stein (2002) a rgues that decentralize d firms are

more likely to undertake soft-i nformation projects, as proje ct managers in these types

of firms have more autonomy to make de cisions. Stein shows this by comparing the

expected outputs of firm divi sions when information is sof t or hard and when firms are

decentralized or hierarch ical. Hierarchical firms undertake hard information projects,

as management is centralize d and relies on the input of several pr oject managers.

Central management selects har d-information projects bec ause hard information is

verifiable. Banking is an excel lent way to test this theory, as , according to Stein,

banking is a well-defined “industry”where: “(a) it is easy to identify the primar y

‘projects’that line manager s must choose among—namel y, individual loan

applications; and, moreove r, (b) it seems quite plausible t hat information about these

particular projects is likel y to be innately soft”(p. 1913). T echnology may allow

information to travel more qu ickly between bank employees , especially in hierarchica l

firms. Namely, this can mean tha t central management has a la rger role in lending

decisions. If this is the case, improv ed technology will have a negative as sociation

with soft-information len ding.

This article is also related to relationship banking, which is defined as a

relationship in which an intermediary both invests in obtaining customer-specific

information (which is often proprietary) and evaluates the profitability of these

investments through multiple interactions with the same customer over time or across

products (Boot 2000). The customer-specific information is often soft information,

which the bank then processes when making lending decisions. A decrease in

soft-information lending is likely to be related to decreases in relationship lending.

1

Below, I modify this ratio slightly to capture employees who are solely working in the bank’s lending area.

6 The Journal of Financial Research