...to a number of research and policy issues.

Introduction

Small businesses are an important part of the economy of virtually every nation. In the United States, for example, small businesses account for about half of all private-sector employment and nonfarm gross domestic product according to the Small Business Administration (SBA). (1) Nonetheless, small firms have historically faced significant difficulties in accessing funding for creditworthy (that is, positive net present value) projects due to lack of credible information about them by potential providers of funds. Small businesses are typically much more informationally opaque than large corporations because they often do not have certified audited financial statements to yield credible financial information on a regular basis. Also, these firms usually do not have publicly traded equity or debt, yielding no market prices or public ratings that might suggest their quality. To address this opacity problem, financial institutions use a number of different lending technologies.

This paper focuses on small business credit scoring (SBCS), a lending technology used by many financial institutions over the last decade to evaluate applicants for "micro credits" under $250,000 ($250K). SBCS involves analyzing consumer data about the owner of the firm and combining it with relatively limited data about the firm itself using statistical methods to predict future credit performance. Credit scores have been widely used for many years in consumer credit markets (for example, mortgages, credit cards, and automobile credits). This has resulted in low-cost, commoditized credits that are often sold into secondary markets, yielding significant growth in consumer credit availability. However, only in the mid-1990s did financial institutions begin to combine the consumer and business information to create scores for small business credits on a widespread basis, and to date, no significant secondary market for small business credits has emerged. Later, we will describe the SBCS technology and its use, review the extant research on its effects on credit availability (broadly defined), and discuss the key research and policy issues related to this technology.

To put SBCS into context, it is one of a number of transactions lending technologies based primarily on "hard" quantitative information used by financial institutions to address the opacity problem. The hard data for transactions technologies are relatively quickly gathered without need for prior contact with the firm and can be relatively easily observed, verified, and transmitted to others. Other transactions technologies using hard information to lend to opaque small businesses include asset-based lending, factoring, fixed-asset lending, and leasing. Each of these technologies is based primarily on a particular source of hard information other than quality financial statements and traded securities, which are not available for opaque small businesses (for example, asset-based lending is based on valuations of accounts receivable and inventory pledged as collateral). (2) For SBCS, the hard information is primarily personal consumer data on the owner obtained from consumer credit bureaus, data on the business collected by the financial institution, and in some cases, information on the firm from commercial credit bureaus. Though SBCS is often used to evaluate opaque small businesses, it may also be used for relatively transparent borrowers to reduce underwriting costs.

As an alternative to the transactions technologies, financial institutions also attack the opacity problem using relationship lending based on "soft" qualitative information gathered through contact over time with the firm, and often with its owners, managers, and other members of the local community. The soft information generally takes significant time to accumulate and is not easily observed, verified, or transmitted to others. Financial institutions' choices among the technologies used to evaluate small business credit applicants vary across both institutions and applicants. These decisions depend on the comparative advantages of the institution, the information available about the firm, and the expected costs and accuracy of each of the feasible technologies, all of which affect the expected profitability.

The remainder of the paper proceeds as follows. The next section discusses the development of SBCS over the past decade. Then a brief description is given of a survey of SBCS usage that is the basis for much of the active research on this topic. The literature concerning the effects of SBCS on small credit availability is then reviewed, after which follows a discussion of research and policy issues related to SBCS and a conclusion.

SBCS

Credit scoring is a statistical approach to predicting the probability that a credit applicant will default or become delinquent. (3) Though this underwriting method is well established in consumer credit markets, it has only been widely applied to small commercial credits for about the last decade. (4) The primary reason for the delayed response is the significant heterogeneity among borrowers (making prediction difficult) and significant variation in the underwriting approaches both within and across banks and for credits of various sizes (reducing the ability of banks to pool data). The change occurred when analysts determined--particularly for the very smallest credits--that credit information for the principal owner explains a significant amount of the variation in the performance of small business credits. This may reflect, in part, a correlation between personal and business success and, in part, a commingling of the finances of the business and the owner. Incorporating this information directly into SBCS models improved prediction and encouraged greater underwriting standardization.

The first SBCS models were constructed by collecting a sample of fully seasoned loans and then dividing them into "goods" and "bads" on the basis of a credit event such as default or delinquency. (5) This binary outcome was then statistically related to a number of characteristics about the principal owner and business that are believed to predict loan performance. With respect to the principal owner, this included credit bureau information on variables such as income, net worth, available credit, prior delinquencies, and prior bankruptcy. Similar information about the business was also used in the statistical model: financial ratios (such as profitability and leverage), the presence of past credit problems (if a business credit report was available from a commercial credit bureau like Dun & Bradstreet [D & B] or Experian), as well as the type of business (standard industrial classification [SIC]). Though as many as 50-60 variables may have been considered when building a model, only 8-12 variables were ultimately used because of significant multicolinearity. Limited dependent variable estimation techniques, such as logit and probit, were then generally used to analyze the data to see how well the variables predicted the credit events. (6)

Though some large banks, such as Wells Fargo, have developed proprietary SBCS models, most institutions have turned to outside vendors. The largest external provider, Fair Isaac and Company (Fair Isaac), introduced its first SBCS model in 1995. The model used a sample of more than 5,000 small business loan applications over five years from 17 large U.S. banks designed to represent a national pool. This model, which was constructed in cooperation with the Robert Morris Associates, was further refined in 1996 using data from 25 banks. Several large banks began to adopt SBCS following the introduction of the first Fair Isaac model.

Today, even lenders with proprietary models often purchase credit scores from outside vendors, especially when making loans outside of communities in which they maintain an office. In the case of Fair Isaac, several models are now available through their Small Business Scoring Service depending on the type of credit (for example, loans, leases, lines of credit), the type of information available (for example, application data, business data, consumer data), and the size of the credit (for example, less than $50K, less than $100K, less than $250K). Although the models have been designed for use for credits up to $250K, some lenders use them only for amounts less than $100K. The firm is currently investigating whether current models can be validated above $250K and is also developing new models for these larger credits.

There may be several motivations for lenders to use SBCS. Cost saving is likely a key incentive. This may be particularly the case when banks purchase scores from outside vendors and use these scores for automated accept or reject decisions and setting credit terms. This implementation of the SBCS...

NOTE: All illustrations and photos have been removed from this article.



More articles from Journal of Small Business Management
Investment patterns of informal investors in the Alberta private equit..., January 01, 2007
Venture capital in Spain by stage of development., January 01, 2007
On government programs that increase small firms' access to capital., January 01, 2007
Board composition in small and medium-sized family firms., January 01, 2007
On governance and agency issues in small firms., January 01, 2007

Looking for additional articles?
Search our database of over 3 million articles.

Looking for more in-depth information on this industry?
Search our complete database of Industry & Market reports by text, subject, publication name or publication date.

About Goliath
Whether you're looking for sales prospects, competitive information, company analysis or best practices in managing your organization, Goliath can help you meet your business needs.

Our extensive business information databases empower business professionals with both the breadth and depth of credible, authoritative information they need to support their business goals. Whether it be strategic planning, sales prospecting, company research or defining management best practices - Goliath is your leading source for accurate information.