Update: Read the analysis of 2016 nonemployer establishment data.

Last week’s release of the new Bureau of Labor Statistics (BLS) Contingent Worker Survey (CWS) has renewed public interest in tracking the number of self-employed or independent contractors. The upcoming release of the U.S. Census nonemployer establishment data for 2016 this Thursday will provide another useful data point about the trends in self-employment.

Our analysis of nonemployer establishment data through 2015 indicates that these data confirm previous general findings regarding Uber drivers, all online demand platform work, and independent contractors more generally. These previous findings are:

1. There is an increase in the amount of people involved in these types of work activity;
2. The increase is primarily among people who are doing so to earn supplementary income and for a short amount of hours. The increase in the various self-employment activities has not occurred in people’s “main job” or as their main source of income; and
3. The economic scale of these activities has not increased greatly when measured as a share of hours worked or compensation earned. While “headcount” measures do show a large increase, the overall economic impact is relatively small.

As we will show, there has been a significant increase in the headcount of nonemployer establishments. However, the economic impact, as measured by the share of nonemployer establishment revenues in total revenues, has not increased in roughly 20 years. This indicates that the growth of nonemployer establishments seems to reflect the growth of self-employed individuals operating unincorporated businesses that generate very little revenue.

Background on CWS and nonemployer data

The CWS is based on a household survey and showed that in 2017, the share of independent contractors in total employment was 6.9 percent, comparable to what was reported in all of the other CWS surveys back to 1995. This has elicited surprise by many observers, especially those that have touted the explosion of gig work or freelancing.

The release of new Census data on nonemployer establishments for 2016 this month provides another opportunity to track these trends. These Census data capture the phenomenon of self-employment by examining business establishment/firm data rather than asking individual workers or households about employment. Census describes the data as:

“data for businesses that have no paid employees and are subject to federal income tax. Most nonemployers are self-employed individuals operating unincorporated businesses (known as sole proprietorships), which may or may not be the owner’s principal source of income.”

Various analysts have used these data to track the rise of self-employment with many arguing that the rapid rise of nonemployer establishments reflect a surge of self-employment and gig work. Abraham et al. (2017) note that the growth of nonemployer establishments was faster than self-employment in the Current Population Survey (CPS). They cite this evidence (along with tax records) as being consistent with the argument that the CPS is missing some of the growth of in self-employment, pointing out that nonemployer establishments “have trended upwards as a percent of the number of earners.” Steve King of Emergent Research, a leading analyst of gig and “independent worker” trends, has also said that nonemployer data are “a useful general indicator of U.S. self-employment.”

All of the analyses we have seen of these data solely focus on headcount measures, tracking the number of nonemployer establishments, and do not examine the economic impact, as we do in our analysis by examining the revenues of nonemployer establishments.

This may be surprising since the Census used to (but now does not) warn data users of the following:

“The majority of all business establishments in the United States are nonemployers, yet these firms average less than 4 percent of all sales and receipts nationally. Due to their small economic impact, these firms are excluded from most other Census Bureau business statistics.”

Nonemployer establishment trends, 1997–2015

Figure A shows the rapid growth of nonemployer establishments since 1997. The data for the Economic Census years (every five years, in years ending in 2 or 7) are noted with red markings to indicate their more reliable trend. The latest data on the number of nonemployer establishments in 2015 were respectively, 12, 47, and 58 percent greater than in 2007, 2000, and 1997. In contrast, the number of firms and establishments with employment grew by much less, respectively, by 6.5 and 11 percent from 1997 to 2015. These trends confirm, as other headcount measures have also shown, that self-employment activity grew far faster than that of regular establishments.

But what is the economic scale and impact of this rise in nonemployer establishments? The Census Bureau provides data on the total receipts (i.e. revenue) of nonemployer establishments each year, enabling us to assess the growth of their economic impact in terms of the amount of revenue and the share of total revenue earned by nonemployer establishments. We adjust revenues by the growth of the Gross Domestic Product (GDP) price index to track trends in “real” (or inflation-adjusted) revenue. Doing so allows us to see that from 2000 to 2015 the number of nonemployer establishments grew far faster, 47 percent, than their total real revenues, up just 20.5 percent. Inflation-adjusted revenues per nonemployer establishment fell by 18.1 percent from 2000 to 2015. These trends suggest that the net growth among nonemployer establishments was among establishments with very low revenues. Again, this evidence is consistent with much self-employment being supplementary side-work rather than a person’s main job.

How much has the economic importance of nonemployer establishments increased since 2000 or from prior to the Great Recession? It turns out that there has been no increase in the economic scale of nonemployer establishments. Figure B shows that nonemployer establishments’ share of total (nominal) revenue has not grown in the 18 years since 1995, maintaining a roughly 3.2 percent share of total revenue.

The table below shows that, in absolute terms, total real revenue grew faster in the overall economy than among nonemployer establishments. In fact, since 2000, real revenue grew more than 50 percent faster in the overall economy, 32.5 percent, than among nonemployer establishments, 20.5 percent. Since 2007, which includes the rapid growth of online platform activity, nonemployer establishment revenue grew by only 2.5 percent, considerably slower than aggregate revenue growth of 8.2 percent.

Taxi industry

The nonemployer establishment data do show a rapid rise in taxi industry nonemployer establishments, presumably reflecting the growth of Uber and Lyft: up from 192,000 in 2011 to 480,000 in 2015. The economic impact of this surge in activity was blunted, however, as the real revenue per taxi nonemployer establishment fell by 34 percent from 2011 to 2015. The nonemployer establishments in the taxi industry have grown to 2.0 percent of all nonemployer establishments by 2015 but generate just 1.0 percent of all nonemployer revenues. It is a familiar story: Lots more activity, not much of an economic impact.

Read more on Uber and the economy.