1 Journal of Banking & Finance 22 (1998) 1067±1076 ``New'' data sources for research on small business nance John D. Wolken * Board of Governors of the Federal Reserve System, Mail Stop 149, Federal Reserve Board, 20th and C Sts. NW, Washington, DC 20551, USA Abstract This paper describes three new sources of data on small business nances: Bank Call Report data on small business lending, the 1995 Survey of Consumer Finances (SCF), and the 1993 National Survey of Small Business Finances (NSSBF). Each of these data sources o ers publicly available micro-level data useful for examining a wide variety of issues and questions about small business nances. A number of studies which have utilized these data are cited and information on how to access these data is provided. Ó 1998 Elsevier Science B.V. All rights reserved. JEL classi cation: D21; G20; M20 Keywords: Small business nance; Small business data; Small business loans 1. Introduction Until recently, up-to-date and comprehensive data on small business nances have been lacking. Some data have been available in aggregate form, such as the Internal Revenue Service's Statistics of Income (1994) and the Bureau of the Census' Characteristics of Small Business Owners (1991). However, data at the micro level generally have been unavailable to the public, primarily due to concerns about con dentiality and privacy. * Tel.: ; /98/$19.00 Ó 1998 Elsevier Science B.V. All rights reserved. PII S ( 9 8 )
2 1068 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067±1076 In the past few years, new sources for studying issues of small business - nance have begun to appear. In this paper, I will describe three important ``new'' sources of publicly available micro-level data which can be used to study a variety of issues in small business nance. These three sources are Call Report data on small business loans, the Survey of Consumer Finances (SCF), and the National Survey of Small Business Finances (NSSBF). 1 For each of these data sources, the discussion includes a brief description of the content, a summary of important advantages and limitations, and a few citations of research using these data. 2. Bank and thrift Call Report data on small business lending Since 1993, the Federal Reserve and other regulatory agencies have required insured depository institutions to report on small business lending in the midyear Reports of Condition and Income (Call Report). 2 These data were collected to measure the extent of insured depository institutions' lending to small businesses. The unit of observation for this data set is a federally insured depository institution. In 1997, there were about 9000 commercial banks and about 2000 thrift institutions reporting. Some reporting problems occurred in the rst year (1993), but since 1994 the data have been reliable (see Table 1). Insured depository institutions report on two types of ``small business'' loans: (l) Commercial and Industrial Loans to US Addresses and (2) Loans Secured by Nonfarm Nonresidential Properties. For each type, depository institutions report the number and amount outstanding for loans with origination amounts of less than $100,000, $100,000±$250,000, and $250,000 to less than $1,000,000. Origination amounts are the maximum of the loan extension, loan commitment, or total loan value if the extension is part of a loan participation. (Similar information is also collected for two types of small farm loans.) Attractive features of the Call Report data set are: (i) it is a census, not a survey ± all insured depository institutions are required to report these data on the Call; (ii) it is fairly timely ± data are available within six months; (iii) panel data can be constructed from each annual cross section to study changes in depository lending to small businesses over time; and (iv) information about the characteristics of the depository institutions (available from the Call Report) can be matched with the small loan data. Unfortunately, these data do 1 For a general survey of small business data sources, see Ou (1991). A compilation of statistics on small businesses from a wide variety of sources is available in the U.S. Small Business Administration's Handbook of Small Business Data (1994). 2 In 1997, the banking agencies required larger banks to report on small business loan extensions as part of the new Community Reinvestment Act provisions. These data will be a useful addition to the small business Call data.
3 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067± Table 1 Thrift and bank small business loan data (schedule RC-C, Part II) Federally insured depository institutions Number of observations About 9000 commercial banks and about 2000 savings institutions in 1997 Collection frequency Annually (since 1993) on June 30th Call Report General description Detailed information about an institution's balance sheet, loan portfolio, deposit portfolio, and income statement Small business related elements Number and amount of small loans outstanding for ``Commercial and Industrial Loans to US Addresses'' and ``Loans Secured by Nonfarm Nonresidential Properties'' for three loan size origination classes: $0±$100,000, $100,000±$250,000 and $250,000±$1,000,000 About six months Call Report information can be purchased from the National Technical Information Service (NTIS). See for more information not include any loan contract information other than loan size nor are data available on the rms to which these loans are made. These data can be (and have been) used to measure the extent of small business lending by depository institutions. To do so, however, requires the assumption that small loans are originated with small rms. In some situations, this assumption may not be tenable. First, not all small loans are originated with small businesses. The extent of this problem is unknown at this time. However, as a result of 1997 revisions to the Community Reinvestment Act, data will be available in 1998 to answer this question. As part of the Community Reinvestment Act, larger banks will be required to report on the number and amount of loan extensions made to businesses with fewer than one million dollars in annual receipts. Second, depository institutions certainly make loans to small businesses that are not booked as commercial and industrial loans or nonfarm nonresidential property loans. Examples include some auto loans and second mortgages on personal residences when the proceeds are used for business purposes. Despite the newness of these data, a sizeable body of literature has used them to examine issues centering around bank consolidation and the e ect of bank size on small business lending. Two examples are the papers in this special issue by Peek and Rosengren (1998) and Strahan and Weston (1998). Two other recent studies examining bank mergers and small business lending are Berger et al. (1998) and Walraven (1997) Survey of Consumer Finances The 1995 SCF is the most recent of a series of surveys of US households conducted about every three years for the Board of Governors. This survey
4 1070 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067±1076 collects information from a nationally representative cross-sectional sample of about 4000 households. The unit of observation is a household. Detailed information about household assets, liabilities, pension and retirement plans, use of nancial services and suppliers, and demographic characteristics were obtained (see Table 2). In particular, the survey contains a section on business ownership and management, including number of businesses owned, number of businesses in which the household has an active management role, the type of business, how and when acquired, number of employees, and percentage of businesses owned. There are also questions about the relationship between business and personal assets. One advantage of this survey is that it is more likely to obtain information about ``home-based'' and start-up small businesses than are surveys of small businesses. This is because the sample design of the SCF calls for oversampling of households with certain wealth characteristics that are likely to be rare in the general population. These same characteristics are likely to be positively correlated with business ownership. In contrast, the smallest and newest businesses are typically underrepresented on publicly available business lists from which business surveys are likely to draw their samples. Another advantage of the SCF is the depth of data collected on the personal assets, liabilities, and nancial services used by the household. These data can be combined with business ownership information to study the relationship between personal and business nances for small rm owners. Additionally, the questions about business ownership have remained virtually unchanged since the 1989 SCF, permitting cross-sectional comparisons over time. Table Survey of Consumer Finances US household Number of observations Over 4000 Collection frequency About every three years General description General survey of the components of household wealth, including assets, liabilities, retirement, demographic characteristics, employment, pensions, and use of nancial services and suppliers Small business related elements Information on active management and ownership of businesses, including numbers of employees, family members working in business, percentage of businesses owned, relationship between business and personal assets, organizational form and business type About two years (but declining) ndex.html. This site contains links to the 1995, 1992, and 1989 SCF surveys, codebooks, related documentation, and technical papers on survey methodology
5 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067± Table National Survey of Small Business Finances US small business ± de ned as for pro t, non nancial, nonfarm business enterprise with fewer than 500 full-time equivalent employees and in operation as of year-end 1992 Number of observations Over 4600 Collection frequency About every ve years General description Second survey of a nationally representative sample of small businesses conducted by Board of Governors and Small Business Administration. The purpose of the 1993 survey was to collect information on small business credit availability, but designed to provide general research database on small business nances Small business related elements Firm and owner demographics, an inventory of nancial service and supplier use, characteristics of nancial institutions, most recent loan experience, trade credit, equity injections, and a balance sheet and income statement About three years (but declining) This site contains links to survey data, codebook, methodology report and some research reports. The 1987 data will be added in the near future To date, I am aware of only one study which has attempted to use these data to study small business nancing questions. In this issue, Avery et al. (1998) examine the role of personal wealth in nancing small businesses. Their study uses data from both the SCF and the NSSBF, which is described next National Survey of Small Business Finances The 1993 NSSBF is the second survey of small business enterprises co-sponsored by the Board of Governors and the US Small Business Administration. 3 The survey collected data for 1993 through interviews conducted with business owners in 1994 and early The unit of observation was a small business enterprise. A cross-sectional sample of 4637 business enterprises were selected to provide a representative sample of all nonfarm, for pro t, non nancial small business enterprises in the United States. These rms represent about 5 million businesses (see Table 3). For the purposes of this survey, small businesses were de ned as enterprises with fewer than 500 full-time equivalent employees that were for pro t, non - 3 The rst NSSBF collected data for rms operating during Many of the questions used in the earlier survey were repeated in the 1993 survey. However, changes in the lists of businesses from which the samples were drawn changed between 1987 and This makes direct comparisons between the 1987 and 1993 data problematic. For such a comparison, see Cole et al. (1996).
6 1072 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067±1076 nancial, nonfarm business enterprises and in operation as of year-end Full-time equivalents were de ned as full-time employees plus one half of part-time employees. Although a 500 employee rm may actually be quite large, using such a broad de nition facilitates comparisons between truly small businesses and ``medium'' sized businesses. The survey solicited information about the characteristics of the rm and its primary owner (for example, rm and owner age, industry, and type of business organization), the rm's income statement and balance sheet, and details of the use and sources of nancial services. The survey also obtained information about the rm's most recent borrowing experience, the use of trade credit, and capital infusions. Most data apply to Balance sheet and income data are for scal year-end The 1993 NSSBF is the most comprehensive source of data available on small businesses' use of nancial services and nancial suppliers. 4 One unique feature of this data set is that individual nancial service suppliers (and their characteristics) can be linked to speci c nancial products used by small businesses. This feature was used by Kwast et al. (1997) to examine the extent to which businesses (and households) ``cluster'' their purchases of nancial services at local nancial institutions. Elliehausen and Wolken (1992) examined the clustering issue for small businesses using the 1987 NSSBF. Petersen and Rajan (1994), and Berger and Udell (1995) are examples of studies that used the 1987 NSSBF to examine the importance of lending relationships in small business lending. In this issue, Cole (1998) examines relationship lending using the 1993 NSSBF. Cole et al. (1996) used both the 1987 and 1993 NSSBFs to determine whether the competition to banks from nonbanks increased over the 1987±1993 period. And as mentioned earlier, Avery et al. (1998) combined data from the SCF and the NSSBF to study the role of personal wealth in small business nance. Issues other than small business lending have also been examined using the NSSBFs. For example, Cole and Mehran (1997) studied the determinants of executive compensation in small businesses. And Ang (1992) used data from the 1987 NSSBF to evaluate a proposed theory of nance for privately held rms. 5. Other sources There are several other sources which provide important information about small rms, but whose micro-level data generally are unavailable to the public. 4 For general statistics on the use of nancial services by small businesses, see Cole and Wolken (1995).
7 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067± Data on loan contract terms for small business loans are rare. One source which has been used to make inferences about small business loan terms is the Federal Reserve's Survey of Terms of Bank Lending for Businesses (STBL). This survey collects loan contract terms for all loan extensions during one week each quarter from a sample of about 300 commercial banks (Table 4). For each commercial and industrial loan extension, the survey collects the loan amount, interest rate, maturity, commitment, whether the loan is part of a participation or is callable, repricing, and purpose of loan. The data are timely, typically available within about six weeks of the reporting period. These data have the same limitations as do the small business loan data on the bank call report. In particular, size of loan must be used as a proxy for size of business and no data on the characteristics of the business are collected. Balance sheet and income statement data for small businesses are often dif- cult to obtain. Most small businesses do not have audited nancial statements and many do not prepare balance sheets or income statements. However, representative nancial statements can be obtained from the Internal Revenue Service's Statistics of Income Book (Table 5). About three years following a tax year, the IRS publishes detailed income statement items for sole proprietorships, corporations, and partnerships. Balance sheet items are published for corporations and partnerships. The data are compiled annually from a large sample of tax return data led by businesses. Within these publications are tabulations for di erent industries, asset and sales size groups, and organizational forms. Information on business ownership characteristics, such as race and gender are generally unavailable, publicly or privately. Every ve years, the Census Bureau conducts economic censuses. One of these surveys is the Characteristics of Business Owners (CBO) survey (Table 6). The data are collected through a statistically selected sample mail survey of about 100,000 business owners using Table 4 Federal Reserve's Survey of Terms of Bank Lending to Businesses Individual bank loan extension Number of observations More than 30,000 loans per quarter Collection frequency Quarterly for one or more days of the rst week of the second month of the quarter General description Detailed information on contract terms of all domestic commercial and industrial loans originated during survey period Small business related elements Data on loan or commitment size, interest rate, loan amount, whether collateral was pledged, whether issued under commitment, purpose of loan, months to maturity About six weeks Statistical Release E2; also published in Federal Reserve Bulletin in February, May, August, and November
8 1074 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067±1076 Table 5 Internal Revenue's statistics of income Business or individual tax return Collection frequency Annually General description Detailed list of balance sheet and income statement items as compiled from tax returns. No balance sheet data for sole proprietorships Small business related elements Published data are reported separately for several receipt, asset, and industry categories, for di erent types of corporations, partnerships, and sole proprietorships About three years for published Book. Some data is available earlier via electronic media Corporate Book of Statistics of Income (Washington, DC: Internal Revenue Service) and SOI Bulletin (Washington, DC: Government Printing O ce). Internet site: administrative records. The 1987 CBO provides data about the characteristics of minority and women-owned businesses and their owners as well as a comparable nonminority, male-owned business universe. Also included in the CBO are questions on owner age, marital status, birthplace, educational background, work experience of owner, household income of owner, and starting capital requirements. Data from the 1992 CBO should be available soon. 6. Summary This paper has described three ``new'' publicly available sources of data that are useful for studying issues in small business nance. In addition, the paper brie y describes three data sets that are somewhat less accessible, but nevertheless provide useful information about the characteristics of small businesses and their loans. The studies referenced in this paper which have used one or Table 6 Bureau of the Census' Characteristics of Business Owners survey Owners of sole proprietorships, small partnerships, and Subchapter S corporations (less than 10 owners) Number of observations About 100,000 business owners Collection frequency Every ve years ending in 2 and 7 Small business related elements Owners attributes, attributes of business startups, business income and employment, sources and amount of startup capital About ve years. (Data for 1992 survey should be available soon.) 1987 Characteristics of Business Owners (Washington, DC: Government Printing O ce, November 1991)
9 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067± more of these sources are testimony to the interest in small business nance. Even though the data have been available for only a short time, they have already proven valuable in gaining a better understanding of small business - nances and in formulating public policy a ecting small businesses. Acknowledgements The opinions expressed do not necessarily re ect those of the Board of Governors or its sta. The author thanks Nicole Meleney for excellent research assistance. References Ang, J.S., On the theory of nance for privately held rms. Journal of Small Business Finance, 1. Avery, R.B., Bostic, R.W., Samolyk, K.A., The role of personal wealth in the evolution of small business nance. Journal of Banking and Finance 22, 1019±1061, this issue. Berger, A.N., Saunders, A., Scalise, J.M., Udell, G.F., The e ects of bank mergers and acquisitions on small business lending. Journal of Financial Economics 50. Berger, A.N., Udell, G.F., Relationship lending and lines of credit in small rm nance. Journal of Business 68. Bureau of the Census Characteristics of Business Owners. US Government Printing O ce, Washington, DC. Cole, R.A., The importance of relationships to the availability of credit. Journal of Banking and Finance 22, 959±977, this issue. Cole, R.A., Mehran, H., Determinants of executive compensation: Evidence from small businesses. Working paper Federal Reserve Board, Washington, DC. Cole, R.A., Wolken, J.D., Financial services used by small businesses: Evidence from the 1993 national survey of small business nances, Federal Reserve Bulletin 81. Cole, R.A., Wolken, J.D., Woodburn, R.L., Bank and nonbank competition for small business credit: Evidence from the 1987 and 1993 national surveys of small business nances. Federal Reserve Bulletin 82. Elliehausen, G.E., Wolken, J.D., Small business clustering of nancial services and the de nition of banking markets for antitrust analysis. The Antitrust Bulletin 37. Internal Revenue Service Statistics of income: Corporate income tax returns, Washington, DC: US Department of the Treasury, Internal Revenue Service. Kwast, M., Starr-McCluer M., Wolken, J.D., Market de nition and antitrust in banking. The Antitrust Bulletin (forthcoming). Ou, C., Available nancial data bases for research on small business. In: Yazdipour, R. (Ed.), Advances in Small Business Finance. Kluwer Academic Publishers, Dordrecht, pp. 169±178. Peek, J., Rosengren, E.S., Bank consolidation and small business lending: It's not just bank size that matters. Journal of Banking and Finance 21. Petersen, M.A., Rajan, R.G., The bene ts of rm-creditor relationships: Evidence from small business data. Journal of Finance 49. Strahan, P.E., Weston, J.P., Small business lending and the changing structure of the banking industry. Journal of Banking and Finance 22, 821±845, this issue.
10 1076 J.D. Wolken / Journal of Banking & Finance 22 (1998) 1067±1076 US Small Business Administration. Handbook of Small Business Data. US Government Printing O ce, Washington, DC. Walraven, Nicholas, Small business lending by banks involved in mergers. Federal Reserve Board Finance and Economics Discussion Series 25.