“Form ever follows function. This is the law.”
This hallowed tenet of modern design was first put forth in the late nineteenth century by Louis Sullivan, Chicago architect and father of the skyscraper. The size and structure of an object, he asserted, should reflect and be dictated by the purpose for which the object is intended. This philosophical shift spawned over a century’s worth of ever sleeker high-rises and sedans, coffee tables and salt shakers. But can size and shape turn the table, take the lead, force function to follow their form? This question is at the heart of recent work by Kellogg School of Management finance professor Mitchell Petersen and colleagues published in the Journal of Financial Economics. The researchers describe how differences in size between banking behemoths and small-town lenders influence the types of strategies the institutions pursue and the types of clients they can best serve.
“A common theme in all my work is to look at the supply of capital. In this case, do lending decisions of big and small banks differ? That’s a great question,” said Petersen, the Glen Vasel Professor of Finance and a Research Associate with the National Bureau of Economic Research. “Why can both big and small banks both exist? If one is more efficient, how does the other survive? Could there be distinct markets amenable to different sizes and structures?”
While a Fortune 500 semiconductor manufacturer may ooze spreadsheets and financial statements, such hard information may be relatively scarce at the corner Laundromat.The nation’s banking industry teemed with roughly 14,000 banks through the 1970s and early 1980s. However, that number started to collapse in the mid- to late-1980s, fueled in part by deregulation, technological innovation, and subsequent consolidation (see Figure 1). One explanation for the decline was that the country had too many banks for an economy of its size, so the saturated banking market needed to shrink. Other economists maintained that small banks were an inefficient technology that gave way to bigger banks that could be more efficient due to economies of scale.
Number of banks, branches, and total offices
Source: FDIC Historical Statistics on Banking, Table CB01
Information, Soft and Hard
Petersen suggests that those explanations do not tell the whole story. “As big banks bought them up, small banks could no longer provide lending to small, soft information firms,” he said. “So other small banks started up in markets where there weren’t enough banks focusing on relationship lending and soft information.”
New banks have continued to proliferate despite an overall decline in the number of banks due to consolidation. The growth trend has been firmly entrenched in the last decade, with the largest number of new charters for FDIC-insured institutions occurring in 2006.
So-called “hard” information is easy to collect, put in a spreadsheet, and pass on to someone else. For example, banks could analyze a company’s record of annual expenditure on wages. But soft information, while often just as useful, can be difficult or even impossible to measure, analyze, and communicate. Petersen highlighted these difficulties when he asked, “When a loan officer asks if a borrower is honest, whose version of honesty do you use? How do you document it?”
The researchers theorized that large banks can function quite efficiently by employing one group of specialists to collect information about potential borrowers and another group to conduct analyses and make lending decisions based on that information. However, this process hinges upon the ability of borrowers to provide hard information like detailed financial records that can be easily measured, communicated, and analyzed throughout the organization. Large banks incentivize employees to focus on collecting and analyzing hard information, the life’s blood of the organization.
Large banks, however, could be hamstrung by businesses that have little to offer in the way of hard information. While a Fortune 500 semiconductor manufacturer may ooze spreadsheets and financial statements, such hard information may be relatively scarce at the corner Laundromat. Many small businesses cannot afford to hire armies of accountants to conduct regular audits and maintain detailed financial records, but many of those same small businesses are unparalleled in their integrity, industry, and ingenuity. While such qualities can forge a business, support a community, and fuel an economy, none would appear in a database.
To make “character” loans to small companies based on soft information, the researchers propose that banks should have the lending decision made by the same employees who collect the information, since these employees would be most familiar with the intangible qualities that borrowers possess. This link between information gathering and decision making often already exists in smaller banks. Employees at smaller banks tend to be incentivized to collect the best information about borrowers that they can, since these same employees play a role in deciding how to allocate bank resources via loans. Smaller banks tend to have a finger on the pulse of their communities since they often focus their funds on the families, businesses, and farms in their vicinity.
“It’s this ability of small firms to get credit for soft information and invest that we argue is an engine of economic growth in this country,” Petersen said. “You should be able to get credit for your good projects, even if you’re a small firm with mostly soft information.”
Bank Size as a Predictor of Lending and Relationships
To understand credit options firms enjoy, or obstacles they face, Petersen and colleagues amassed and dissected data on firms, banks, and the relationships they share. They used the Federal Reserve’s 1993 National Survey of Small Business Finance to identify 1,131 small, for-profit firms, each of which employed less than 500 people. The median firm reported assets worth $680,000, with the top 25 percent reporting at least $2.85 million and the bottom 25 percent no more than $150,000. Only 57 percent of the firms kept detailed accounting records or financial statements.
The researchers then identified banks that lent to those firms during the years 1990 to 1994. Information was drawn from the FDIC Summary of Deposits and from Consolidated Reports of Condition and Income—commonly known as Call Reports—collected by the Federal Financial Institutions Examination Council. The median asset size of the banks was $956 million, with the smallest 25 percent having no more than $163 million, and the top 25 percent having at least $7.69 billion. The size of the average loan those banks made to the small firms was $1 million; the smallest 25 percent were no larger than $125,000, and the largest 25 percent were at least $600,000.
Petersen and colleagues explored relationships between firms, banks, and loans using a mathematical modeling technique called ordinary least squares regression. They determined which features, such as business size and loan amount, could be combined and calculated to best predict, for example, bank size. The model parameters were adjusted until the calculated predictions best matched the actual values.
“We’re essentially saying that, by definition, soft information is hard to measure and computerize, yet we computerized it and ran regressions,” said Petersen. “We have indirect measures of soft information, such as the size of the firm. It would be nice to have more direct measures, or additional indirect measures, of soft information.”
As Petersen and his colleagues had predicted, size mattered: larger banks typically focused their lending on larger firms. Compared to a smaller bank, a bank that was 40 percent larger would typically make loans that were twice as large—to borrowers that were twice as large. Further supporting the theory, larger banks were more likely to focus on firms that could provide hard information. Borrowers who maintained detailed financial records and could provide this hard information often received loans from firms that were 24 percent larger than those firms that loaned money to borrowers supplying primarily soft information.
The ability (or desire) to conduct business over coffee and a handshake was also influenced by bank size. Firms’ relationships with larger banks were less personal, relying more on telephone and mail rather than on face-to-face interactions. This tendency to teleconference rather than power lunch was also reflected by increased distance between large banks’ branches and the borrowers they served. The distance between bank branch and borrower more than doubled, and the probability of having impersonal relationships increased from 15 percent to 38 percent, as the size of the bank increased from the 25th percentile ($163 million) to the 75th percentile ($7.69 billion).
Bank size also influenced the stability of relationships with borrowers. Older, more exclusive relationships with borrowers are more prevalent when banks are smaller. As bank size increased from the 25th to the 75th percentile, the probability that clients borrowed from just one bank plummeted from 74 percent to 27 percent, and the length of the relationship with the average firm was cut almost in half, from 8.8 years down to 4.5 years.
Small Firms and Large Banks
These findings lend clear support to aspects of the researchers’ theories on size and structure, information and function. But a fundamental, practical question remained: Is it harder for small businesses to obtain loans from larger banks? To find out, the team first had to devise a way to identify credit barriers that businesses face.
Small firms, which depend on a small bank’s willingness to lend based on soft information, may be forced to borrow from large banks if there are not many small banks in a particular market. But large banks, designed to prefer larger, hard information firms, may manage risk when lending to small firms by limiting their access to credit, keeping them on a short, tight leash. Feeling pinched, small firms may then have to turn to more costly sources of credit. One such alternative source is trade credit from their suppliers.
Petersen argued that constrained credit from larger banks leads small businesses to be later in repaying trade credit to their suppliers. Analyses of trade credit repayment among 546 small businesses supported this theory. An increase in the size of the bank—from the 25th to 75th percentile—increased the percentage of trade credit that borrowers paid late, from 26 percent to 43 percent. Some businesses clearly encountered limited access to credit as a result of the mismatch between small business size and large bank size.
“For hard information and big firms, it’s easy to document a good credit risk. But what’s the supply of credit from big banks when there are no incentives for dealing with small firms?” asked Petersen. “This kind of information may affect not only how much credit you get, but also whether you get it at all, and from where.”
In addition to highlighting characteristics of large and small businesses and banks in the United States, these findings could help guide development efforts around the globe. Said Petersen, “When large banks or government agencies go into developing countries, they naturally want to lend based on hard information. But instead of direct lending, maybe they could invest in small, local, community-type lenders, small enterprises on the ground that know who’s who, know the kind of information that can’t easily be passed up the chain of command.”
“This doesn’t argue that big banks are bad,” said Petersen, adding that “both kinds are filling niches.” Although this yin yang dynamic may not inspire Confucian introspection in the traditional sense, it could help individuals discern a funding path in quests like building a gas station by the interstate or buying a deep fryer for the mom and pop doughnut shop.
Sullivan, Louis H., “The tall office building artistically considered,” Lippincott’s Magazine, March 1896.
Jack of All Trades or Master of One?
Performance perceptions of multi-feature products