Are US firms getting larger?

Some people say that the market is increasingly dominated by large companies. Well is that true? I would like to know to. So I dug up some data on enterprises that the US Census Bureau keeps to try to shed some light into this.

Granted, I am not too sure if this data set alone would paint the perfect picture, but in the world of the blind, the one-eyed man is king. With that in mind, I’m not going to dive into the specific numbers to the n-th decimal place, with more of an emphasis on the big picture here. Do take this write up with a grain of salt and do your own research. The data is available here: https://www.census.gov/programs-surveys/susb.html

So let’s plot things out. Here is a plot of the breakdown of the proportion of US businesses by number of employees:

Figure 1: Distribution of US firms by number of employees (1997–2017). Looks very constant and flat to me at this level.

Well, this graph doesn’t tell much — the distribution looks very flat and constant. Here are the figures for 2017 the latest year available:

• 89.04% of firms have fewer than 20 employees;
• 10.62% of firms have 20–499 employees;
• 0.17% of firms have 500–999 employees;
• 0.10% of firms have 1,000–2,499 employees; and
• only 0.07% of firms have 2,500 employees or more.

If we plot out the distribution of US employees by firm size, we might see a tiny little suggestion from the graph that the proportion of employees working for a large 2,500+ firm is marginally increasing over time:

Figure 2: Distribution of US employees by firm size (1997–2017). If you squint hard enough, you’ll see that the proportion of employees working for firms with more than 2,500 employees marginally increasing.

Here are some numbers to give the graph some perspective — in 1997, 19.11% of employees were working for a firm with fewer than 20 people while 36.04% were working for a firm with 2,500 employees or more; in 2017, those numbers are 16.41% (down ~3 percentage points) and 40.43% (up ~4 percentage points) respectively.

Now if we plot out the distribution of receipts (the money coming in) by firm size, we do see a similar small increase in the amount received by larger firms:

Figure 3: The distribution of receipts by firm size in the US (1997–2017).

In 1997, firms with fewer than 20 employees had 15.28% or receipts whereas firms with 2,500 employees or more had 47.44%; in 2012, the numbers were 12.24% (down ~3 percentage points) and 51.11% (up ~4 percentage points) respectively. In other words, receipts were roughly in line with the distribution of employment.

So, the next question one could possibly ask is are larger companies a good thing? Well, I don’t know if this data set can answer that. And I guess the definition of a good thing can be subjective. The only thing that I can plot out from here is the implied productivity per employee, i.e. I have divided the receipts in each bucket by the number of employees in that bucket. The data does show that employees are consistently more productive at larger firms.

Figure 4: The implied productivity per employee by firm size in the US (1997–2012).

The figure does show a clear stratification of productivity levels — the larger the firm by employee numbers, the more productive each employee is. In 1997, firms with fewer than 20 employees brought in ~$138,000 per employee whereas in firms with 2,500 employees or more that number was ~$228,000 per employee; in 2012, those numbers are ~$195,000 (up 41%) and ~$368,000 (up 61%) respectively. This also suggests that productivity growth is larger in firms with more employees.

Conclusion: These numbers do suggest that by a small amount, large firms are having more of an impact in terms of the proportion of firms, people employed, and receipts. But will this trend continue? Or would firms become smaller as technology allows firms to be nimbler? I don’t know. Only time would tell.

Here’s one last chart to sum everything up, plotting the growing significance of large firms:

Figure 5: Summary — plotting the growing significance of large firms. Note: vertical axes for this plot do not start at 0.