What's Behind High Small-Biz Failure Rates?
Flawed statistics may make entrepreneurs' prospects look worse than they are
Q: Do you have any idea where I can locate some statistics that reflect the failure rates of high-tech startups? For instance, I'd like to know how many businesses are started each year and how many survive the startup stage.
--B.B., San Francisco
A: Numerous studies have looked at small-business failure rates over the past 30 years. Unfortunately, the results are often contradictory or vague. Getting good data on small, privately held companies -- and interpreting it accurately -- isn't easy because they don't have to report their financial results.
William Dennis, senior research fellow for the National Federation of Independent Business' Education Foundation, analyzes a survey of 36,000 households each year for the Wells Fargo/NFIB Series on Business Starts and Stops. The latest one, available on the NFIB's Web site, www.nfibonline.com, reported that 2.9 million businesses were launched in 1997, involving nearly 4 million people. Another 1 million people purchased 700,000 existing businesses that year.
The data showed that most new businesses are very small. More than two-thirds start in the owner's home, and only 21% initially employ someone besides the owner, says Dennis, who will release the 1998 report shortly. He doesn't track startups by industry category.
For business terminations, the Wells Fargo/NFIB study uses data of the U.S. Census Bureau, which only records closures of companies with employees. Those statistics show that about half of businesses that employ people are still operating five years after they open. "I feel good about the accuracy of the startup numbers," Dennis says, "but there are undoubtedly a lot of underreported stops."
The NFIB estimates that over the lifetime of a business, 39% are profitable, 30% break even, and 30% lose money, with 1% falling in the "unable to determine" category. Even when a business closes its doors, there can be many reasons for what's statistically a "failure," including a sale or merger, which may actually be a sign of robust financial health or good prospects. "When a business ends, it may be because the investors have lost their investments or because they have sold out profitably," Dennis notes.
The premise that there are many reasons -- not all of them bad -- for business closures, was behind a study by John Watson and Jim E. Everett published in the Journal of Small Business Management in October, 1996. The authors studied 5,196 startups in 51 managed shopping centers across Australia between 1961 and 1990 to try to determine true failure rates, because they felt that bankers and venture capitalists were basing negative views of small companies on dubious statistics.
Their data showed that annual failure rates were greater than 9% when failure was defined simply as "discontinuance of ownership." When failure was defined as bankruptcy, however, the number dropped to less than 1% annually. About 4% of the businesses that closed their doors each year "failed to make a go of it." And owners disposed of about 2% of businesses annually to prevent further loss. The authors concluded that cumulatively 64.2% of the businesses failed in a 10-year period -- if failure was measured as discontinuance of ownership -- but only 5.3% actually filed for bankruptcy during a decade.
For more information, check the U.S. Census Bureau's Characteristics of Business Owners Database at www.census.gov. Several commercial research firms collect financial data on businesses, compile it as reference and research material, and sell it to lenders and other interested parties. Other sources of small-business data are: Dun & Bradstreet, www.dnb.com; Robert Morris Associates, www.rmahq.com; and Financial Research Associates, www.frafssb.com.
Have a question about running your business? Ask our small-business experts. Send us an E-mail at email@example.com, or write to Smart Answers, BW Online, 46th Floor, 1221 Avenue of the Americas, New York, NY 10020. Please include your real name and phone number in case we need more information; only your initials and city will be printed. Because of the volume of mail, we won't be able to respond to all questions personally.