Moonlighting Mystery: The Missing Multiple Job-Holder

Multiple job holders are the stuff of urban economic legend. Yet they fail to show up in government statistics. Turns out it may be because of the way questions are asked and statistics compiled. Income below a certain level is frequently classified as a 'hobby,' although it certainly doesnt feel like that to those who feel forced to supplement their incomes. The sharing economy, as a result, is less about sharing than about surviving. JL

Justin Fox reports in Bloomberg:

Job benefits and job security have been squeezed over the past few decades. Far less apparent in the data, though, is any dramatic shift to a"gig economy" or a land of freelancers. Yet.
In a column about an interesting proposal for a new approach to employment benefits, Felix Salmon referred to "a large increase in the number of people working multiple jobs, for well over 40 hours per week in total, but not receiving any of the benefits which have historically accompanied that kind of workload."
That seemed like it might be right, but I've been finding a lot of surprises in employment data lately. So I went to the Bureau of Labor Statistics' website, and found this:

The numbers are from the Current Population Survey, the monthly survey of about 60,000 households from which the unemployment rate is derived. While multiple jobholding is down overall, the subset of these multiple jobholders who have only part-time jobs has grown -- from 1.2 percent of employment in 1994 to 1.3 percent today. So yes, an increase, but not exactly a large one.
Maybe the problem, though, is with the questions. If you have a part-time job and make a few thousand extra bucks a year selling needle-felted miniature animals on Etsy, are you going to tell a CPS surveyor that you have multiple jobs? I know I wouldn't say that, yet I file a Schedule C with the Internal Revenue Service every year to report book royalties, speaking fees and freelance writing income.
Happily, there is another data series that does capture such work. It's assembled by the Census Bureau from IRS tax return data, it comes out once a year with a two-year delay, and hardly anybody pays any attention to it. I only discovered that the most recent numbers were released late last month because I was looking at old posts on Steve King's excellent Small Business Labs blog.
The Census Bureau's "nonemployer businesses" are sole proprietorships, partnerships or corporations that have no employees but report $1,000 or more in annual receipts -- unless they're in construction, in which case the cutoff is $1. In the chart below I also include the BLS data series on self-employment, which is derived from the Current Population Survey and only goes back to 2000 because that's when the government started keeping track of the incorporated self-employed along with the unincorporated self-employed. If every nonemployer business were a full-time job, the two lines would presumably overlap. They don't -- so the gap roughly represents people who are getting business income on the side.

Given how different the data sources are (tax returns versus a survey), this is only a rough approximation. But the gap is definitely growing. Also, the Census Bureau actually tightened its screens for identifying nonemployer businesses starting in 2009, meaning that the gap has probably been growing even faster than the chart shows.
In 2013, there were about 8 million more nonemployer businesses than people reporting that they were self-employed. This is a lot of people. There's no telling how many of them actually face the predicament described by Felix Salmon at the beginning of this column, but it's got to be in the millions given the sharp increase during the Great Recession in people working part-time because they couldn't get full-time work (as I wrote last week). What this data doesn't show, though, is some kind of sweeping transformation of the American workplace, in a nation of 149 million employees as of May.
If you want to see what a sweeping transformation of the American workplace looks like, consider these two numbers on retirement plans: In 1982, 84 percent of full-time workers at medium-sized and large companies were enrolled in defined-benefit pension plans; by 2013 it was down to 27 percent. This is clear evidence of "The Great Risk Shift" that political scientist Jacob S. Hacker wrote about a decade ago and that has left Americans less financially secure than they used to be. In general, job benefits and job security have been squeezed over the past few decades. Far less apparent in the data, though, is any dramatic shift to a"gig economy" or a land of freelancers. Yet.
The U.S. Government Accountability Office made some minor waves this spring by reporting that "contingent workers" (part-timers, temps, contract workers and the self-employed) now make up 40.4 percent of the workforce. That was up from 30.6 percent the last time the GAO made such an accounting, in 2006, which seems like a pretty dramatic leap. But if you look more closely at the report you see that it's an apples-to-oranges comparison.
The data source used in 2006 was a Bureau of Labor Statistics survey conducted in 2005 after similar surveys in 1999 and 1995. Those surveys, interestingly, showed a decrease between 1995 and 2005 in contingent workers' share of overall employment. Basically, contingent work has always been a big part of employment in the U.S. It's not a new phenomenon, although it has been moving from agriculture and construction to more white-collar fields.
Since 2005, the BLS hasn't been able to get funding to do another such survey (thanks, Congress!), so the GAO relied instead on the General Social Survey conducted by the University of Chicago's NORC (which at some point in the past stood for National Opinion Research Center). The slightly different accounting of contingent work that the GAO culled from the GSS (sorry about all the acronyms) came to 35.3 percent of employment in 2006 and 40.4 percent in 2010 -- still an increase, although not as dramatic as the purported leap from 30.6 percent to 40.4 percent. Also, the GSS survey sample is much smaller, and thus subject to a greater margin of error.
The contingent-work increase from 2006 to 2010 consisted almost entirely of a rise in the percentage of part-time workers -- not too surprising, given that the economy was only just beginning to recover in 2010.
A new set of GSS results from 2014 came in too late for the GAO to include in its report. I couldn't figure out how to reconstruct the GAO's full contingent-work estimate from them, but I was able to see that the percentage of part-time workers plummeted between 2010 and 2014. So it's hard to say if the 2006-2010 increase was evidence of a secular trend or just a really bad recession.
If you've been stuck in part-time jobs for the past seven years while wanting to work full-time, you probably don't care if this is the result of secular or cyclical forces. Also, it may well be that we are on the cusp of dramatic change in the nature of work as technology destroys jobs and possibly enables new kinds of independent and semi-independent work. There are certainly a lot of people in Silicon Valley and elsewhere betting on this. We just haven't gotten there yet.

1. In case you're wondering, now that I’m a Bloomberg employee the speaking fees are off the table and the freelance income will be infinitesimal. Let's keep those book royalties coming, though.
2. The reasoning here is that businesses with receipts below $1,000 are deemed hobbies, and I guess the Census Bureau figures nobody would do paid construction work as a hobby.