Do small companies actually create more jobs? Maybe not

For many years, the accepted wisdom has been that most jobs are created in smaller companies. But newly released figures from the Bureau of Labor Statistics in the US challenge that notion.

“The most growth in employment has been in large firms,” said Nathan Clausen, the bureau’s economist in charge of the development of the new statistics, which were released this month after more than two years of development.

The figures cover employment from April 1990 — one month after employment reached a high for that economic cycle — through March 2011, just over a year after employment hit bottom after the 2007-9 recession.

Over that entire period, employment at large companies — defined as those having at least 500 employees — rose 29 percent, while employment at smaller companies rose by less than half as much.

At small companies, defined as those with fewer than 49 employees, the total number of jobs rose by just 10.5 percent over the period. Had the entire private sector grown at that pace, rather than rising by more than 19 percent, as it actually did, there would have been nearly eight million fewer people with jobs last March, and the unemployment rate would have been around 14 percent.

The new data was released on an experimental basis, and is shown in the accompanying charts. If all goes as planned, in July the bureau will release figures through March of this year. Within a couple of years the figures will be released with only a three-month delay and will also include information on wages and hours. That will enable analysts to see which size companies pay better, both over all and in various industries, and where pay is rising or falling.

Small-company employment seems to have been more stable in good times and bad. The numbers rise more slowly during expansions and fall less sharply during recessions. When large companies shed workers during and after the 2001 recession, the number of people working for small companies actually rose.

When the data becomes available on a monthly basis, it may help to provide indications of cyclical change. At economic turning points, small companies seem to be more nimble. In June 2007, small companies began to reduce employment, seven months before larger ones did. Small companies began to add workers in November 2009, four months before larger ones did.

The bureau also released numbers for broad industry groups, which showed substantial variations. In manufacturing, smaller companies seem to have done much better than larger ones.

Employment at smaller companies fell at a slower pace during the crisis, and since then has risen more rapidly. Seasonal variations also can vary substantially. In retailing, an area where the growing dominance of larger employers has been particularly notable, nearly all the seasonal variation comes at larger companies, as they hire Christmas-season workers. But in construction, smaller companies are much more likely to let people go during seasonal lulls than are larger ones.

So where did the belief that small companies create most jobs come from? To a large extent, it comes from another data series, known as Business Employment Dynamics, maintained by a different part of the Bureau of Labor Statistics.

For each quarter, it calculates the number of jobs added and lost in each size of company, and those figures seem to show that over the long term larger companies create about 35 percent of jobs, as opposed to more than 60 percent in the new numbers.

The differences may relate to how companies are classified by size. The new data classifies companies based on their maximum employment during a year. That minimizes how often companies change categories.

A ski resort, with a lot of employees in the winter and only a handful in the summer, might be considered a large company based on its peak employment. In the other data, that company would change size every season. There are also minor differences in which workers are included and excluded.

But it turns out the Business Employment Dynamics data is quite similar in one respect. It also shows that the proportion of private sector employees working for large companies has been steadily rising in recent decades.