The Atlanta Fed's macroblog provides commentary on economic topics including monetary policy, macroeconomic developments, financial issues and Southeast regional trends.

Authors for macroblog are Dave Altig and other Atlanta Fed economists.

« What Is the Stance of Monetary Policy? | Main | Thinking About Progress in the Labor Market »

February 26, 2014

The Pattern of Job Creation and Destruction by Firm Age and Size

A recent Wall Street Journal blog post caught our attention. In particular, the following claim:

It’s not size that matters—at least when it comes to job creation. The age of the company is a bigger factor.

This observation is something we have also been thinking a lot about over the past few years (see for example, here, here, and here).

The following chart shows the average job-creation rate of expanding firms and the average job-destruction rates of shrinking firms from 1987 to 2011, broken out by various age and size categories:

In the chart, the colors represent age categories, and the sizes of the dot represent size categories. So, for example, the biggest blue dot in the far northeast quadrant shows the average rate of job creation and destruction for firms that are very young and very large. The tiny blue dot in the far east region of the chart represents the average rate of job creation and destruction for firms that are very young and very small. If an age-size dot is above the 45-degree line, then average net job creation of that firm size-age combination is positive—that is, more jobs are created than destroyed at those firms. (Note that the chart excludes firms less than one year old because, by definition in the data, they can have only job creation.)

The chart shows two things. First, the rate of job creation and destruction tends to decline with firm age. Younger firms of all sizes tend to have higher job-creation (and job-destruction) rates than their older counterparts. That is, the blue dots tend to lie above the green dots, and the green dots tend to be above the orange dots.

The second feature is that the rate of job creation at larger firms of all ages tends to exceed the rate of job destruction, whereas small firms tend to destroy more jobs than they create, on net. That is, the larger dots tend to lie above the 45-degree line, but the smaller dots are below the 45-degree line.

As pointed out in the WSJ blog post and by others (see, for example, work by the Kauffman Foundation here and here), once you control for firm size, firm age is the more important factor when measuring the rate of job creation. However, young firms are more dynamic in general, with rapid net growth balanced against a very high failure rate. (See this paper by John Haltiwanger for more on this up-or-out dynamic.) Apart from new firms, it seems that the combination of youth (between one and ten years old) and size (more than 250 employees) has tended to yield the highest rate of net job creation.

John RobertsonBy John Robertson, a vice president and senior economist in the Atlanta Fed’s research department, and

Ellyn TerryEllyn Terry, a senior economic analyst in the Atlanta Fed's research department

February 26, 2014 in Economic conditions, Employment, Labor Markets, Small Business | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference The Pattern of Job Creation and Destruction by Firm Age and Size:


So young high growth firms that succeed create the most jobs. WOW!

Large successful old firms are stable and neither create nor destroy large amounts of new jobs. The analytical insight is unbelievable!

and to top if off small, risky start-ups destroy the most jobs as they constantly fail. OMG, Nobel Price of Economics right there!

Americans please send even more of your tax dollars to the Atlanta Federal Reserve given the amazing level of research and analytical insights they are capable of.

Posted by: Alex | February 27, 2014 at 01:46 AM

Are the highlighted features of the chart -- that "rates of job creation and destruction tend to decline with firm age" and that "the larger dots tend to lie above the 45-degree line, but the smaller dots are below the 45-degree line" -- stable over the time period examined, or do they come and go from year to year? And if the latter, can that change be correlated in a useful way with events in the economy as a whole (eg the "dot bomb" of 2001, the financial crisis of 2007-8, or the Great Recession to which that crisis gave rise)?

Posted by: Ed Blachman | February 27, 2014 at 09:29 AM

I keep thinking about the pattern and I wonder what it would look like if it was in motion through time. I wonder if it would follow patterns in nature. I also wonder if the large industries 4 years or less move around much since it is so unusual for a company to hire that many employees in such a short time because of outliers. I also wonder the implications for future employment levels as there are fewer entrepreneurs. Thank you so much for posting this. Great article. I've read other stuff you've done that you linked out to. Great job guys. You've found a lot of the critical data points that really matter. I've been looking to find something like this. I will bookmark this page.

Posted by: buttmunch1 | February 28, 2014 at 12:17 AM

One question: Would the chart look much different if you stopped at 2006? The idea is that large numbers of small firm jobs were likely destroyed by the financial crisis. Further, after the crisis small firms have lacked access to start-up funds, thus diminishing gross new firm creation (and its attendant jobs). One can think of the GFC as a seminal event in small firm birth/death dynamics, much more so than for large firms. While it may look from the data that small firms don't have much impact on net job creation, this may be because of the lack of small firm "births" in the past five years.

Posted by: Diego Espinosa | March 07, 2014 at 12:17 PM

The chart is better art than economics.

The obvious correlation between firm age and firm size means that it is impossible to separate the effects of the two factors (on job creation) by assigning colors and sizes to firms and graphing them against each other. The "cure" for multicollinearity is not changing the color or size of data points -- but recognizing that both change simultaneously.

A successful startup firm in 1987 moved along a path from then to 2011 that took them from tiny blue dots in the direction of giant orange balls. What does that PATH suggest about job creation and destruction? We can be certain that companies traveling the same path in the opposite direction have a far higher rates of job destruction to creation. Shall we paint both of them green, and assign both medium-size dots?

Finally, I was confused by the artist's practice of measuring dependent variables along both principal axes, then graphing observations for independent variables as points in the x-y plane. This is perfectly fine if the sole purpose is to describe the data in a compact way. But if one's purpose is to guide the reader's mind toward cause-effect relationships, this is a poor practice.

Posted by: Thomas Wyrick | March 08, 2014 at 10:05 AM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search

Recent Posts

November 2014

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29



Powered by TypePad