An Analysis of Small Business Size and Rate of Discontinuance
基于1987年美国人口普查局的企业主特征调查数据，发现小企业停业率并非如Birch所言那样高，且停业率与企业规模呈反比。
(Tables and figures not included) David Birch's conclusion that firms employing fewer than twenty workers have created about 88 percent of all net new jobs in 1980s is wellknown to scholars in small business field. Challenged by some .and quoted by many, it may be said to epitomize recent resurgence of interest in small business. A second conclusion of Birch's is somewhat less familiar. He argues that aggregate growth in employment forthcoming from the great American job creation machinethe nation's small business sectorhas at its base great volatility, a foundation of massive, continual failure. If this is indeed case, then volatility of small business community may preclude possibility of sustained employment for much of growing portion of labor force that works in this sector. THE STUDY AND DATABASE This premise of massive, continual small business failure-held by many others besides Birch-is examined critically in our present study. Among small firms with paid employees, principal finding is that discontinuance (or-broadly interpreted-failure) rates are not massive. Among small firms with five to nineteen employees, for example, annual rates of business discontinuance were found to be only about 3 percent. Among largest small business employers, annual discontinuance rates were less than 1 percent during period from 1982 to 1986, inclusive. The measured rate of discontinuance among small firms is highly dependent upon definition of what is or is not small business. Discontinuance rate estimates are presented in this article for broadly defined small business universe as well as for enterprise subsets that are defined by gross receipts categories and employment. Discontinuance rates are found to be inversely related to firm size, whether is measured by employment levels or grass sales revenues. The Characteristics of Business Owners (CBO) survey, compiled by U.S. Bureau of Census in 1987, is used in this study to estimate discontinuance rates for small business universe and various enterprise subsets. According to IRS data, about 12 million proprietorships, partnerships, and small business corporations existed in 1982. The CBO survey was drawn from this universe. In August 1986, questionnaires covering both owner and business traits were sent out to 125,000 persons who owned businesses in 1982. This survey produced an 81 percent response rate, and completed questionnaires provided basis for constructing CBO database, empirical basis of present study. In following analysis, CBO data are weighted by (1) legal form of organization, (2) industry mix, (3) geographic location, and (4) owner's demographic traits, in order to be representative of 1982 small business universe of proprietorships, partnerships, and small business corporations. FINDINGS Table 1 is based upon owner responses to CBO survey question, Is business you owned in 1982 still operating? Businesses are defined as discontinued if owner's response was no. In cases where multiple owners of one firm existed, one response per firm was permitted for purposes of this study. More than nine thousand sample observations were dropped because they represented multiple owners of partnerships or corporations. Table 1 indicates that 34.0 percent of small businesses in existence during 1982 had discontinued operations by late 1986. Among firms reporting 1982 sales of less than $5,000, discontinuance rate was 49.3 percent, while at other end of size spectrum, only 8.2 percent of firms with sales exceeding $1 million had discontinued operations. Table 1 clearly indicates that discontinuance rates are inversely related to firm size, with very small businesses accounting for bulk of 1982 universe that had discontinued business operations by late 1986. In table 2, discontinuance rates are summarized for small business employer subset of CBO database. …