Home Mail Articles Stats/current Supplements Subscriptions Links
The following articles appeared in Left Business Observer #66, October 1994. They were written by Doug Henwood, editor and publisher. They retain their copyright and may not be reprinted or redistributed in any form - print, electronic, facsimile, anything - without the permission of LBO.
Two cliches of American economic life are that small business is the engine of job growth and that corporate downsizing is an indispensable tonic for productivity. Two studies published this past spring by the U.S. Census Bureau's Center for Economic Studies (CES) show that the dispensers of these nostrums should check their facts.
The small business myth is probably the most durable and pervasive of all. It holds appeal across the political spectrum, from corporate lobbyists trying to sell tax breaks to postmodern New Agers trying to sell their vision of decentralization and local self-reliance. We are constantly told that universal health care, or a higher minimum wage, or tighter environmental regulations will drive the small business job machine off the rails. But whatever the virtues of intimate scale, like friendliness, informality, the possibility of cooperative or democratic governance, from a purely economic point of view, small isn't beautiful. Small firms pay less than large ones, are less likely to offer health, pension, or child care benefits, and are often more dangerous to workers. With few exceptions, they're not all that innovative technologically. And now it emerges that in manufacturing at least they are not the job machines they are reflexively praised for being.
Aggregate employment data like national and even local statistics on job growth and loss provide only a crude picture of the labor market. Behind the smooth curves of business cycle upswings and downswings or the waxing and waning of geographic regions lie a tremendous rush of job creation and destruction. One of the first attempts to examine that fine picture in detail is Gross Job Flows in U.S. Manufacturing, a book-length study from the CES by Steven Davis, John Haltiwanger, and Scott Schuh. They draw on the Census Bureau's vast store of information on some 400,000 U.S. manufacturing plants (the Census Bureau counts businesses, governments, and houses as well as people). Their repudiation of the small business job machine myth is a by-product of their investigation of the birth and death of manufacturing jobs between 1972 and 1988.
Probably the most striking feature of the study is the incredible volatility of the labor market. In a typical year of that 16-year span, about one in ten manufacturing jobs disappeared, and as many were created -- even though the total number of manufacturing jobs was virtually unchanged over the period at a bit over 19 million. Most of these changes are persistent and concentrated, that is, once a job in a particular plant disappears, it stays disappeared, and both growth and decline tend to be concentrated at plants that are either growing or shrinking markedly. If you add together the workers in all industries, not just manufacturing, who lose jobs, change jobs, or get their first jobs, 37% of the labor force changes its employment status every year -- really a staggering rate of turnover. While such volatility might seem more typical of the U.S. economy than Europe's, the evidence suggests that rapid turnover is the norm in all rich industrial economies.
Of great political interest are the authors' findings on where the jobs are and aren't. On balance, new jobs do not sprout in the greatest numbers at either fresh startups or small firms. In their words, "large, mature plants and firms account for most newly created (and newly destroyed) jobs." Smaller employers do generate plenty of jobs, but they also destroy them in great quantities. If you add together creation and destruction, no clear picture emerges. As the study says, "In a nutshell, net job creation... exhibits no strong or simple relationship to employer size."
Now it may be that small service businesses do a better job than manufacturers, but there's no good way of knowing; nothing like the Census of Manufactures exists for the services. But the myth of small business fecundity was developed from bad interpretations of bad data. Here are some of the more egregious errors.
* Most analysts define small businesses as those with 500 employees or fewer -- not exactly small in some eyes, but that's not the point. If a firm with 600 employees has a bad year and shrinks to 400, this will show up in bad studies as a growth of 400 jobs in the small business sector, not a loss of 200 jobs from the bigger .
* Many counts ignore job death, and just focus on new hires.
* The database many analysts use, from Dun and Bradstreet, while useful for many commercial purposes, is worse than useless for employment analysis. Audits of the D&B's file by several outside analysts show massive discrepancies not only with official data, which is based on employers' state unemployment insurance filings, and even the phone book. Until recently, the Small Business Administration used the D&B's files for its own internal analyses, and the leading publicist of small business, David Birch of the consulting firm Cognetics Inc., still does.
Birch is the font of much of the "data" on which the small business case rests. As Dan Cordtz reports in the April 26, 1994, issue of FW (ne Financial World), in the late 1970s, Birch bought the D&B's computer tapes and interrogated them in an effort to find out where jobs came from and where they went. In a 1979 report for the U.S. Department of Commerce, he declared that 80% of the jobs created between 1969 and 1976 were produced by businesses with fewer than 100 workers, and nearly two-thirds by firms with fewer than 20. Birch's assertion was picked up by apostles of entrepreneurship like George Gilder and the high-tech worshippers once known as the Atari Democrats.
Myths about firm age travel in partnership with ones about size. Davis, Haltiwanger, and Schuh looked at the relation of job birth and death to plant age, and found that while young plants do hire lots of new workers, they also shed them rapidly; older plants hire fewer workers, but also fire fewer. And smaller units offer jobs with shorter lifespans than larger ones. Older, larger plants, then, offer more job security than young, small ones.
Besides demolishing myths about job creation, Davis, Haltiwanger, and Schuh demolish one about job loss, too. To study the effects of trade on employment, the authors sorted all the industry groups represented into ten classes, five ranked by the ratio of imports to total production, and five ranked by exports. There appears to be little correlation between an industry's exposure to foreign trade, either imports or exports, and employment changes, except for those industries with what the authors call a "very high import ratio," where job loss was three times that seen in the other four categories But these import-exposed sectors account for under 18% of manufacturing employment, and the ranking by export share shows no discernible pattern. So maybe the fetish of export-boosting is as overdone as the small-biz fetish.
Another study, by Martin Neil Baily, Eric Bartelsman, and John Haltiwanger ("Downsizing and Productivity Growth: Myth or Reality," CES Working Paper 94-4) takes on another fashionable doctrine, especially popular on Wall Street, that the road to higher productivity and, eventually, better living for all, is paved with mass layoffs. Also working with the Census Bureau's manufacturing plant database, Baily, Bartelsman, and Haltiwanger find that there's more than one route to higher productivity, and the downsizing route can also be a dead-end.
In this study, the authors divide 140,000 manufacturing plants into four groups, those that both added workers and improved productivity (successful upsizers), those that added workers and lost productivity (unsuccessful upsizers), those that shed workers and improved productivity (successful downsizers), and those that lost both employment and productivity (unsuccessful downsizers). They found that successful upsizers contributed almost as much to aggregate productivity growth as successful downsizers, and that many of the downsizers were really sick or dying rather than gaining leanness and meanness. Successful upsizers also had a growth rate in value added -- the value of output less the costs of inputs like raw materials -- over eight times that of the successful downsizers. The successful downsizers grew, but at a far-less-than-impressive rate.
In line with the Davis & Co. study, Baily & Co. find no great virtue in smallness. The smallest plants, those with under 20 employees, showed a negligible contribution to total productivity growth, and the next range, those with from 20 to 49 employees did only slightly better. The best gains were at plants with from 250-2,499 employees, mainly because of gains at the successful downsizers. The bulk of the successful upsizers, however, were the very largest plants, those with over 5,000 employees, and, perversely, the small businesses that added the most employees were those found among the productivity losers, the unsuccessful upsizers.
Also perversely, the authors could find no neat way of explaining why firms fell into the category they did: size, industry group, region, ownership (that is whether the plant was a lone site or part of a larger, multiplant enterprise), and wage level all failed to provide any clue. Success, at least in productivity, appears to be the result of "idiosyncratic" reasons, Baily, Bartelsman, and Haltiwanger conclude. This is very frustrating news for theorists.
While one might suspect that the U.S. economy labor and labor market are more dynamic or volatile (the word choice depends on your ideological bias) than those of Europe or Canada, a recent study from the Organisation for Economic Cooperation and Development (OECD, the Paris-based official think tank sponsored by the rich industrial countries) confirms the principal findings of Davis, Haltiwanger, and Schuh for the supposedly stodgy outside world. Despite their sclerotic image, the European economies create jobs at roughly the same rate as the U.S., as do Canada and New Zealand. According to the study, published in the OECD's latest Employment Outlook (July 1994), the reason that employment growth is slower in Europe than the U.S. is mainly because jobs there are destroyed at a quicker pace than here (though, of course, international comparisons must be made with the usual cautions about noncomparable data). Also, the rate of job turnover -- the percent of the workforce changing jobs every year -- is roughly the same in Europe, New Zealand, and Canada as the U.S. The standard image that European economies lack dynamism and coddle failure is simply at odds with the facts. Or, if you want to look at it a different way, the incredible volatility of modern capitalism is not just a feature of the U.S. economy, but is seen around the world.
Also, the European experience confirms that the standard picture, which holds that new companies and small businesses are the major source of job growth, is also at odds with the facts. Expansion of existing firms is a far more fruitful source of new employment than startups. And small businesses in most countries are, like the U.S., potent job destroyers as well as job creators. So policies designed to encourage small startups are not likely to boost the rate of job growth.
Another finding from the OECD study is that economic development strategies must be coordinated, not piecemeal. Most government programs in the U.S. focus on isolated businesses -- like New York City giving tax breaks to NBC or the commodities exchange to stay in town. But those businesses have done well in New York not in isolation, but because they are tied to deeply established media and financial businesses also in New York. Similar things can be said about other regions. Believe it or not, the OECD finding is a finding in favor of economic planning, that supposedly discredited idea. Enterprises need links to other enter-prises, to universities and other sources of technical knowledge, and to a workforce gifted with relevant skills. Sink or swim tempered with piecemeal ad hoc assistance -- the American way -- just doesn't cut it.
Home Mail Articles Stats/current Supplements Subscriptions Links