Small Business 2001

Want to do a census of small business? Easy -- so long as you know what you mean by small and business.

Try this the next time you're out with a bunch of company owners: bet them a round of beers that they can't tell you how many small businesses there are in the United States. Our own bet: you'll hear wildly different answers.

Even people who know the numbers are likely to come out with grossly disparate totals, if only because the word business is subject to so many interpretations. Do you mean just "real" companies, with regular offices and employees on the payroll? Are you counting home-based enterprises? What about part-time operations?

For most of the last century even the federal government couldn't have told you how many companies there were. Every five years the Census Bureau counted business enterprises. But it missed a lot of home-based companies, and in any event, the survey has been discontinued. Meanwhile, the Bureau's County Business Patterns, which tabulates the number of establishments in each county, didn't distinguish between independent companies and branches of larger corporations. The Internal Revenue Service's annual count of business tax returns included returns from virtually every legitimate enterprise but also listed returns from a large but unknown number of entities that nobody would really think of as operating businesses, including dummy corporations, condo associations, and so forth. The result: confusion.

In the late 1980s the statistical situation finally began to improve, thanks mostly to the U.S. Small Business Administration. Before then researcher David Birch had been using business records from Dun & Bradstreet to do his well-known job-generation studies. (D&B maintains what is undoubtedly the largest private database of business records.) Then the SBA started leasing D&B's records and began assembling its own database of companies. Eventually, it contracted with the Census Bureau to prepare a full database. That database, known as the Statistics of U.S. Business (SUSB), contains the most current information about companies that have employees.

Then there are all those pesky home-based and sole-proprietor operations, which aren't part of the SUSB (except for home enterprises that have workers). Many of those businesses are "caught" by the Labor Department's regular surveys of people who are self-employed. But such surveys also poll many small companies with employees, and those businesses have already been included in the SUSB. So we can't really get an accurate total-business number just by adding the self-employed to the number of companies with employees.

Bottom line? We'll probably never know exactly how many legitimate businesses there are, let alone how many under-the-table enterprises we'd have to add in to make the statistics really complete. Never mind. Take a walk through the statistics anyway. They aren't perfect or unambiguous. But at least we know more than we used to -- possibly even enough for a friendly wager or two.

How many small companies are out there?
What's a business? A stroll through the statistical thicket.

1. Start with the number of nonfarm companies that had at least one employee in addition to the owner sometime during the year. The latest figure: 5.8 million companies with employees in 1999.

2. Weed out the big and midsize companies.
Fortunately, there aren't very many of those. In 1997 about 16,000 companies had more than 500 employees; 80,000 or so companies had 100 to 500 employees. Those numbers have probably grown a little since then, so say there are 100,000 companies with 100 or more employees; that leaves 5.7 million companies with fewer than 100 employees.

3. Now count the soloists.
Since many small companies have no employees (just the owner or proprietor), next we have to add all the one-person shops. How do we count them? Well, the Labor Department and the Census Bureau conduct periodic surveys asking people if they are self-employed. The response to that question adds up to a pretty big number: 12.8 million self-employed individuals in 2000.

But the self-employed count is also a pretty murky number. First, some people who run their own businesses have employees, so there's a significant overlap between "self-employed" and "companies with employees." Second, the number excludes more than 2 million people who were working for a wage when the survey was conducted but were self-employed at another time during the year. Should their operations count as small businesses, too?

4. So just to make sure you haven't missed anybody ...
Take all the tax returns filed with the IRS that showed some business income. The IRS says it received 24.8 million business tax returns in 1999.

And there may be 25 million-plus business tax returns for 2000 -- the count isn't in. That is the granddaddy of all business-census numbers. Trouble is, those tax returns also include a lot of enterprises that nobody would really think of as a company, like a lawyer who teaches a couple of workshops on the side and reports the income on Schedule C.

5. And the wizard says ...
Since nobody knows the final figure, why not get an estimate? And who better to do the estimate than Bruce D. Phillips, former research director of the SBA's Office of Advocacy and the man largely responsible for putting together all those databases that finally enabled us to count at least some companies? How many real small companies are there -- that is, enterprises with or without employees, with or without an office, that you or we or pretty much anybody else would nevertheless look at and say, yeah, that's a business? Phillips won't commit to an exact answer ("It depends on how you count them," he says), but his final estimate: 15 million to 17 million.

The 2001 State of Small Business issue

Please e-mail your comments to