It would be easy -- when admiring the sky on a summer day -- to think of a far-off thundercloud as languid, majestic, even serene. But as pilots and weathermen know, it would be wrong. Inside that cloud is chaos. Huge updrafts and down drafts flow within and around it. Outwardly stable and inwardly tumultuous, it is the perfect symbol of our New Economy.

Viewed from afar, our economy too looks like a mammoth and billowy cloud, changing shape and size only slowly. Total numbers for job creation and productivity scarcely shift. Inside, though, it is a turbulent, swarming mass of companies, coming and going, rising and falling at extraordinary rates. These flows of business fortune within the U.S. economy raise a host of questions for all of us who are part of them. They constitute an almost continuous changing of the guard.

To gain an accurate picture of this activity inside the economy, we at Massachusetts Institute of Technology and now at Cognetics have been tracking the movements of about 12 million individual business establishments since 1969. We put a "tag" on each establishment (as you might tag a migrating duck) and track what happens to the establishment every six months as it moves about in the "cloud."

There is no simple way we can diagram the entire employment picture, but our economic thundercloud (see figure 1) shows some aspects of it in aggregate terms for 1982 through 1986. The arrows show how jobs move -- as establishments open, close, prosper, or decline -- into and out of the cloud and among its layers.

The thundercloud information and the rest of our data suggest that several aspects of these job flows are remarkable. First, the turnover of jobs within each layer of the cloud is very high. By 1986, almost 43% of the workers in the pool of 20- to 99-employee companies held jobs that were not there in 1982 (see figure 2).

Second, our research also shows that the declines just about offset the additions in most layers. The share of workers -- and, approximately, of the gross national product -- in each layer thus stays remarkably the same even though the flows through each layer are very large. This aggregate stability is why the size and shape of the cloud do not appear to change much when viewed from the outside, even though all hell is breaking loose inside.

Third, most of the flows in the cloud are due to a relatively few companies. As we tagged businesses and followed them, we realized that most companies change very little, and that a few change a great deal. Slow-growing companies (as defined by a growth index that considers both absolute and percentage growth in jobs) make up 85% of the economy, but create just 16% of new jobs. The fastest growers, which represent only 7% of all establishments, create 67%. What's more, most of the significant job creators are companies taking off from the bottom layers of the cloud. Sixty-one percent of the exceptionally fast growers have fewer than 20 employees, and 85% have fewer than 100. By contrast, only 6% of fast growers are companies from the top two pools, which include all companies with 500 or more employees.

We also realized that, on balance, small companies were creating far more jobs than they were losing, while larger concerns were breaking even or losing jobs:



0-19 74%

20-99 35

100-499 -11

500+ 2

In short, what all these statistics describe is a real changing of the guard. Hundreds of thousands of companies are now starting each year. A relatively few of them (mostly small ones) stumble onto something and take off. As they do, they offset the losses occurring as other businesses -- many of them quite large -- lose their innovative edge and decline.

This same pattern of flows can be observed within an industry and within a local economy. Take, for example, the paper industry -- habitually considered a stagnant, gradually declining part of the economy. In fact, while it may be declining, stagnant it is not. As figure 3 demonstrates, a striking number of jobs are being created by new companies forming and smaller companies growing. Companies closing or in decline offset the job gains of their dynamic competitors, but it is clear that what looks stagnant from the outside is actually bubbling with innovation and job growth on the inside.

The same can be said of places. For instance, we all know that not much is happening in, say, Chicago. Or do we? Business activity there from 1982 to 1986 resulted in a net loss of about 100,000 jobs, a relatively small change in terms of Chicago's total labor market. Underlying the net totals, however, were events far more volatile. Thousands of growing establishments created about 900,000 new jobs during the period, offsetting the losses of thousands of declining ones that lost 1 million workers. The net effect: a small aggregate change but a huge changing of the guard.

The magnitude of these flows within the U.S. economy, or its parts, compels a number of observations:

* Working in so rapidly changing a business climate can be hazardous to one's mental health. For this kind of change to take place, one in 5 people has to leave his or her job each year. One in 10 is forced to leave a job involuntarily. One in 10 must make a career change each year. The levels of dislocation are extraordinary.

* Selling to businesses in this economy can be a major challenge. Large accounts can vanish overnight. Finding 20 smaller companies to replace the lost large one is a tough job -- particularly when the smaller ones are embedded in 7 million or 8 million businesses that have no interest in what you have to offer.

* Starting up a business has never been easier -- or more perplexing. As the huge numbers for annual start-ups suggest, the possbilities are almost endless, and the barriers to entry are in most cases minimal. Virtually unlimited freedom raises the difficult questions: where should I start? How do I pick an industry and a place to begin? Where are the odds best, and where are they worst?

* Investing in business is becoming more perplexing, too. Only about 5% of the companies in the United States are publicly held and offer public information about their affairs. Many of the best opportunities are not listed on any stock exchange -- and don't want to be. Most of the growth in the American economy is very private. How do you put your investment dollars into this kind of a system? How do you find -- and share in the growth of -- hundreds of thousands of rapidly growing companies that have no connection with traditional capital markets?

Predicting the future of a turbulent, transitional economy has pushed conventional economics beyond its limits. Many of us need to know where the economy is heading, and are frustrated by the recent track record of economic forecasters. In view of what we have learned about the dynamics of the economy, it should come as no surprise that forecasting is tough. The U.S. statistical system does an excellent job of measuring the parts of the economy that are declining (mostly large manufacturing firms) but records much less about the activities of the smaller, private, mostly service firms that are creating all the growth. We shouldn't expect to be able to predict something we can't even measure. But it is frightening to know how much we don't know. We are flying a jet airplane with Piper Cub instruments.

Governing the unknown is no easier than working in it, selling to it, entering it, investing in it, or predicting it. We observe a host of politicians who take credit when things go well in their jurisdictions and blame the Japanese when things don't. The reality, of course, is that neither the politicians' activities nor those of the Japanese have very much to do with local economic prosperity these days. They have to do with how active or inactive a locale's thundercloud happens to be.

In the months to come, I will share with you what our view inside the thundercloud tells us about finding and keeping a job, reaching customers for your product or service, starting a company with the greatest odds of success, identifying good investment opportunities, and anticipating the future in general.

However sluggish the U.S. economy looks from the outside, on the inside it is anything but. Figure 1 suggests how jobs appear, vanish, or migrate among layers of the economy as businesses experience their fates. Turbulence is everywhere, despite relatively stagnant employment figures overall.

Figure 2 is a close-up of job activity among firms with 20 to 99 employees -- one layer of the cloud. Its numbers are cross-referenced to the thundercloud and provide a key for interpreting other layers. Though the layers are organized by company size (and jobs move among layers as companies move) the creation and elimination of individual jobs has been tracked for each establishment. An establishment is any place people work -- be it a dentist's office, a store, or a branch plant of a larger corporation. The size of a company is considered to be the cumulative size of all its establishments.