What America Needs Is A Few Good Failures
If Arp Instruments Inc. hadn't gone bankrupt in 1981, David Friend probably wouldn't have helped start Computer Pictures Corp., which, in its second year, generated profits on sales of around $6.5 million. If they hadn't started Computer Pictures, Friend and his co-founders couldn't have sold it to Cullinet Software for $14 million last fall. And without the sale to Cullinet, Friend wouldn't be the well-heeled venture capitalist he is today, with two more start-ups under way.
David Birch would make the unsettling argument that the more Arp-like failures an economy experiences, the healthier that economy is likely to be. If that sounds farfetched, consider the Arp case more closely.
Arp was founded in 1969 by Alan R. Pearlman, an engineer with one successful start-up to his credit already. With $100,000 of his own money and $100,000 more tapped from friends and family, he developed the company's first electronic music synthesizer. Soon he was joined by Friend, a 21-year-old graduate student with a background in engineering and music but not in business. The third principal was Lewis G. Pollock, a Boston attorney
This unlikely threesome rapidly built Arp's sales to $7 million and its work force to nearly 200. The company turned in several profitable years. Its products advanced the state of the art in electronic music synthesis. The country's biggest rock stars made music on Arp's high-tech instruments, while the company itself became a star in its own right.
Yet, by 1981, Arp was bankrupt.
Whose fault the bankruptcy was is still an issue of contention (See INC. November 1982 page 38). But it is unarguable that Arp failed ultimately because it could not compete: The particular combination of human and tangible assets that constituated Arp Instruments was not efficient enough to permit the company to remain in business. However, freed by bankruptcy and liquidation, Arp's assets -- some of them, anyway -- have found more valuable employment.
So, in the end, Arp's failure was healthy for David Friend. It was also healthy for the economies of Massachusetts -- where Arp and Computer Pictures were based -- and of the country. In place of one ailing business, several thriving or promising companies now exist. Which is precisely David Birch's point.
Birch, who heads Massachusetts Institute of Technology's Program on Neighborhood and Regional Change, achieved stardom of a sort himself four years ago when he made what was at the time a startling pronouncement Small companies, not corporate giants, create most of the country's jobs Small companies, those with fewer than 100 workers, employ 30% to 40% of the U.S. work force. But, Birch found (through computer analysis of data collected by Dun & Bradstreet on nearly 6 million companies), these small companies generated 82% of the jobs created from 1969 through 1976.
At first it seemed like a paradox: From year to year, decade to decade, the small business share of employment remained about the same, 40% or less. How could small companies generate all those jobs and yet not become a bigger and bigger portion of the economy? That is the wrong way to look at it, Birch says, citing another example to explain why. Divide the population into two groups, he says, children and adults. As individuals, children gain more weight than adults. But, the weight of the children's group remains about the same in proportion to the weight of the adult group. That is obviously because full-grown children enter the adult group. Likewise, small companies that have already completed most of their job generation move into thc big-business category.
If Birch's findings no longer seem especially remarkable, it is because they have been so quickly assimilated into the conventional wisdom. Organizations and individuals who promote the political and economic cause of small business in Washington and in state capitals couldn't quote Birch often enough Credible, academic, and ideologically neutral evidence that small companies generate more jobs than large companies, in proportion to their share of the work force, was valuable ammunition for small business lobbyists competing with other interest groups for attention and favors from Congress and local legislatures.So frequently were they cited that Birch's findings led to thc coining of what Candee Harris of The Brookings Institution in Washington, D.C., calls the Eleventh Commandment: "Thou shalt aid and protect small business because it creates jobs." Birch found himself a much-sought-after luncheon speaker.
Birch's sparkling credentials got a bit tarnished last year when Harris and other members of a research team at Brookings released new, apparently conflicting, statistics about job creation, based on their own analysis of Dun & Bradstreet data. Their work produced numbers not nearly as impressive as Birch's. Small companies, they said, really create only about 40% of the new jobs in the United States, not much more than you would expect from their proportion of the work force.
National Journal's Robert J. Samuelson, an economic journalist widely read and respected by Washington politicians and power brokers, publicized the Brookings study and took potshots at Birch's work. The damage was apparent almost immediately to small business advocates who had been leaning heavily on the Birch revelations to support their political cases. For example, the campaign staffs of two Democratic Presidential candidates, Ohio senator John Glenn and former Florida governor Reubin Askew, got nervous. "After the Brookings study came out," says Jere Glover, a Washington attorney and active Democratic advocate for small business, "we started having problems with them. Askew's staff had been talking to us about a strong small business plank in their campaign, but the next time we called, his policy guy said, 'Hey, these numbers you gave us are in dispute now. They're not clean.' Before, they were hot to trot. Suddenly, they weren't." It took some explaining, Glover says, but most candidates are back in the fold now.
The Small Business Administration, which had commissioned the Brookings research, took much of the heat from lobbyists, angry that the political safety of the Birch findings had been jeopardized by the agency that was supposed to promote small business in Washington. Says one veteran, "In Washington, it's not so important what the truth is, but what people's perception of the truth is." The Brookings numbers, he says, attacked the perception of truth surrounding the Birch findings.
While all this was going on, however, Birch and the Brookings researchers went back to their data and their computers and had another look at the situation. They found that while they still had disagreements, the differences were more technical than real.
The data they both used were collected by Dun & Bradstreet to provide credit reports. They were never intended, Birch says, to be a statistically accurate portrait of anything. To make the data usable, he and the Brookings team had to make assumptions about errors and omissions in the raw numbers. For example, nearly half of the companies in the D&B file that have more than one plant site don't report employment for all of their establishments. So, if the enterprise is credited with only the employees it does report, it appears smaller than it actually is. Birch declined to correct for this understatement and thus credited some job-creation to companies that looked small but were in fact larger. He agrees now that this had the effect of introducing a slight bias in favor of small companies. Brookings researchers, on the other hand, now agree that their counting technique erred in the other direction, giving large companies more credit than they deserved.
Eventually, each side accepted some of the other's logic, and the differences began to narrow. In a more recent look at the years 1976 through 1980, for example, Brookings says that small companies generated 51% of the new jobs -- more than the researchers' original 41% estimate and less than Birch's 82%. "So who is right?" Harris asked rhetorically in a recent seminar. The answer, she said, is that it depends on what time you are talking about and how you measure the small business share. "It is not possible," Harris said, "to say 'the small business share of jobs generated is X' and accept it as gospel."
The number, Birch agrees, is "not nearly so significant as you might think." The proportion of new jobs attributable to small companies in any specific period depends on how many jobs big business creates. When the economy is expanding and large companies are hiring, the ratio declines. When large companies are laying off, small companies can, in fact, be credited with more than 100% of the net new jobs generated. Birch and the Brookings team agree that this ability of small companies to generate new jobs in good times and in bad makes them an invaluable part of the U.S. economy. Having said that, though, the two research groups point to another finding that won't please everyone in the small business community.
Most small companies, 85% to 90% of them Birch and Brookings agree, go through long periods when they don't create a single job.
Their findings have been confirmed by independent research at the University of California at Berkeley. Bruce Kirchhoff, a professor of general business and managerial economics at the University of Nebraska at Omaha, labels the 85% "economic core businesses . . . the low-growth and low-innovation" companies that constitute the vast majority of small businesses. Others call them "lifestyle" companies.
Whether they are service or manufacturing companies, they tend to be in commodity businesses in which profit margins are low but stable. These companies provide a living for their owners. They help hold neighborhoods intact. They provide continuity between generations. They are the mom-and-pop businesses that every politician has hailed hundreds of times. The usual phrase is, "Small business is the backbone of America."
However, they are not the job creators of America.
President Reagan's 1983 report to Congress on the state of small business draws heavily on the work of Birch and Brookings. With a candor that is uncharacteristic of political documents edited so as to avoid offending anyone, the report labels as "myth" the idea that "all small firms generate new jobs. This is not true. A fraction of small firms creates all new jobs created by small firms."
Marjorie Odle and Catherine Armington, Brookings's chief researchers on job generation, found that nearly two-thirds of existing businesses don't contribute positively or negatively to job generation. Rather, the researchers say, economic growth is the result of almost no change in most companies but "violent growth (and shrinkage) in a minority of businesses."
Consider just one box in a mathematical model Birch created to illustrate the job-generation process among companies of different sizes. Inside the box are all the jobs that exist in companies with fewer than 20 employees. Over a two-year period the numberof jobs in the box grew by a modest 2.3%, but Birch wanted to illustrate where the growth came from.
Quite a lot happened over the two years, more than the small net growth in the number of jobs would suggest. Start-up companies added 9.5% more jobs to the box. Declining companies with work forces that fell below 20 employees during the two-year period added 1.4% more jobs to the box. The net contribution of companies that started and ended the period with fewer than 20 workers added another 2.6%. Those were all pluses. On the negative side, 2.3% of the jobs left the box as companies grew larger than 20 employees. Business failures accounted for another 9% job loss.
The point to note from Birch's model is that the gains and losses from company births and deaths, respectively, are nearly four times as large as the resulting 2.3% net gain. Or, put another way, it takes a lot of activity at the margins to create just a little change. "It's like looking at a clock face and seeing the minute hand move. What you don't see," Birch says, "is all the gears turning and the action on the inside."
That is just about as far in understanding the job-generation process as the statistical research at MIT and Brookings has taken us. But if the researchers have drawn a reasonably accurate portrait of that process, and we now know something about how the economy creates jobs, can't we help it along a bit?
One thing the Birch/Brookings work says is that the companies that will be creating many, or possibly most, of the new jobs the United States will need 20 years from now probably don't yet exist. That suggests that one of the things we need to encourage is business births.
Americans hold an important advantage over many of their international competitors in this respect, says Birch, because entrepreneurism is, or has been at any rate, a respected part of the domestic culture. He tells a story of a young engineer, in business for himself, whom he met on an airplane. Birch noticed his accent and asked him why he, an Englishman, had started his consulting business in the United States. "As I was finishing college," the man replied, "I went to meet my fiancee's parents. Eventually they got around to asking me what I intended to do. I said that I was going into business for myself. 'Oh, dear!' my future mother-in-law said, 'Oh, dear!' Her husband was a surveyor for the county. That was when I decided to go to the U.S. to start my business."
As helpful as cultural approbation may be, creating new companies still requires money. If, as is typically the case, roughly 600,000 companies are started annually in the United States, and assuming the average capital required is only $25,000, then the $1.8 billion invested last year by venture capitalists met less than 15% of the start-up needs. The rest of the start-ups, says Birch, are being funded "by everv grandmother and uncle in town."
Maybe they aren't all grandmothers, but most of them are individual investors who have investment alternatives available to them. Tax laws that improve the attractiveness of research and development limited partnerships, real estate investments, oil and gas tax shelters, or the stock of smokestack or high-technology industries may be good tax laws for those businesses, but they add nothing to the pool of seed capital available for start-ups.
Karl Vesper of the University of Washington contends in his recent book, Entrepreneurship and National Policy, that "Government attention to entrepreneurul, new ventures is relatively liitle compared to that given small, ongoing businesses." Both are important, Vesper acknowledges, but entrepreneurship and small business have different needs. Entrepreneurial policies help create an environment that permits someone with an idea to turn it into a business. In other words, they help "outs" become "ins." Small business policies, on the other hand, are there to help "ins" stay in. Neither policy should be ignored.
Small business has lots of people arguing its cause in Washington and in state capitals, because people who are already in business are visible -- members of Congress know who they are and how many jobs their companies represent in their districts. One question Vesper raises is, who lobbies for the entrepreneur, the person who hasn't started a business yet, but might?
Efforts by individuals like Vesper and organizations like the Washington-based Corporation for Enterprise Development to help policymakers draw a useful distinction between entrepreneurial companies and small businesses have borne some fruit. Robert Friedman, CfED's president, says state governments have been the first to recognize that start-ups and established small companies have different needs. He points out as examples of this recognition the creation of development banks and the changing of regulations restricting investment by state pension funds -- both of which improve the supply of seed capital. Legislation that would permit taxpayers to write off their investments in small start-up companies has been introduced in Congress, but no action has yet taken place. Nationally, Friedman says, there is still no organized "constituency" for entrepreneurism. "My worst fear," he adds, "is that someone will say, 'Entrepreneurism! You're right. We need more of that, so let's put another $100 million into the Small Business Administration."
Business births were the largest single factor in Birch's model of the job-creation process. The second largest, business failure, was only a tad smaller. Wouldn't we have more jobs if fewer companies failed?. Again the issue is both cultural and financial.
A British acquaintance recently suggested that one difference between his country and the United States is that here "you're not considered to be a real entrepreneur until you've been through your first bankruptcy. In Britain, on the other hand, bankruptcy marks you for life. Better to have been born illegitimate."
Still, even in this country, no one wants to fail, and again the business "ins" hold an advantage over the unorganized "outs" in their ability to lobby for legislation -- tax relief, tariff laws, regulations -- that would tend to keep them in when market forces might otherwise tend to push them out. Not too many people are in the legislative halls arguing in favor of failure, even when it might have the effect of freeing up assets currently managed by some identifiable "in" for more efficient use by some unknown business "out."
The implications of Birch's revelations about the importance of a highly energized minority of small, young companies to the job-generation process and of the distinction that Vesper draws between entrepreneurs and the rest of the business universe are worrisome when put up against the kind of national policy planning that Republicans and Democrats are talking about today.
Democrats are experiencing an ideological evolution. The old party, created during Franklin D. Roosevelt's era and headed today by House Speaker Thomas P. "Tip" O'Neill Jr., divided the world into two groups -- workers and bosses -- and set the two against each other. The new generation of Democrats has a different vision: Workers and bosses, their differences reconciled, will be joined by government, and together all three will plan for the economic future. They call it a national industrial policy, and while different versions still compete for acceptance, the centerpiece of each of them is a forum where business, labor and government can meet to plan for their common well being Reagan Republicans reject the Democratic idea of creating a national industrial policy because, they say, it offends their free-market sensibilities. More likely it is because they are offended by the idea of giving labor and government a seat at the planning table. They prefer to let business, through its Washington representatives, dictate economic planning. What more protection could any Democrat-inspired industry-labor-government committee have granted to a single company, motorcycle-maker Harley-Davidson Motor Co., than the Reagan White House provided with the 49.4% tariff wall it erected around the ill-managed company's non-competitive product? Whether formally planned or not, this "free market" administration skewed the 1981 tax bill, the biggest business tax cut in history, toward large, capital-intensive industries at the expense of those business sectors in which most employment and market growth was taking place.
None of this is to reject or applaud either the Democratic or the Republican approaches to national industrial planning, but only to put a point on one of the non-quantitative observations Birch makes, which is that the interest groups that affect decision making in Washington, and in state capitals, whether formally or informally, speak for the past, or at best the present -- but never the future. "Rifles," says Birch, "are well represented in Washington. Artificial intelligence is not."
In Holland, business failure is not tolerated. If a large business is financially weak, the government frequently becomes a bail-out investor. Consequently, the Dutch government ends up owning as much as 25% to 30% of these so-called private sector companies. The result, Birch says, is that considerably fewer businesses fail, but the whole Dutch economy is experiencing difficulty in adjusting to new forms of economic growth. It is a tradeoff: micro-security for macro-risk.
The U.S. economy, on the other hand, is still strong -- it has macro-security -- in part because we tolerate a relatively high level of micro-risk: a few companies succeed, most fail, and entrepreneurs are, so far at least, at liberty to try again.
What if the U.S. government had bailed out Arp Instruments? Protected it from foreign competition? It would have saved one company and 200 jobs. And the cost? Well, what is one entrepreneur like David Friend worth? Flow many jobs will the three companies he has started since Arp eventually create? More or less than 200?
We can't be sure, you see. And neither can public-policy makers.
If government uses tax policy and import tariffs to keep old-line American steel companies alive, isn't it also keeping assets out of the hands of entrepreneurs who might, if they had the chance, create the successor industry? What Birch has shown us about the job-generation process suggests, at least, that somebody should be asking the question.
Spokesmen for small business and big business, for manufacturers and retailers, for high tech and low tech, for all the business "ins" -- can always argue that their respective interest group is an important part of the economy and deserves attention. But if Birch's work has taught us anything, it is that safeguarding the untidy process that includes business births, deaths, growth, and decline should take precedence over the preservation of any single interest group. In evaluating public policy, the question should not be, "How many jobs will this save?" Rather we should ask, "Will this foster enterprise?" -- that is, if we are interested in the long-term macro-security of the United States. Unfortunately, that question is not likely to be asked so long as micro-security is what gets votes.