Contingency: a follow-up

My post on contingent and “alternative” work (and the demographic follow-up) annoyed some people who think the Bureau of Labor Statistics, the source of the data, is missing the point through bad definitions and bad techniques. (As am I, for using it.) According to these critics, asking people whether they expect their jobs to last the year is using the wrong definition of contingency—though it’s not clear what the right one is, since most employed people in the U.S. can be fired for no reason at all at any time. Or the BLS was wrong to focus only on the primary job; plenty of people do gig work on the side to supplement their incomes, so the Bureau is—perhaps intentionally!—lowballing the numbers. Yes, though we don’t know how many such people there are (though we’ll know more when the BLS releases its data on “individuals who found short tasks or jobs through a mobile app or website and were paid through the same app or website” on September 30). But the side hustle is a different story from the canonical gig economy line, which is that we’re all day laborers now. Or, best criticism of all, the numbers are just wrong because they don’t comport with my correspondents’ experience. But you don’t know how representative your experience is, do you?

Here are some stats to address these counterclaims.

job tenure

The gig economy story is closely related to idea that the permanent job is a thing of the past—we all bounce from job to job now. That perception is not confirmed by the BLS’s job tenure numbers, which report the median number of years people have been in their current position. Here’s what it those look like like broken down by sex.

Job tenure

Median tenure has bounced around some but the lines are surprisingly trendless, given the chatter about increased volatility. Tenure overall barely changed between 1998 and 2016; it fell by about a year through 1980s and 1990s and rose since; it was about one month longer in 2016 than it was in 1983. That average does obscure some gender differences. Tenure for men fell by a year between 1983 and 1998, rose by half a year through 2014, and then fell a bit between then and 2016, to a level about eight months shorter than 1983. Tenure for women, however, has been rising fairly steadily; it rose about ten months between 1983 and 2016.

There’s not all that much change in tenure by age or industry either. For younger workers, those between 25 and 34, tenure fell by three to four months between 1983 and 2016, with not much variation in the intervening years. By industry, tenure rose by two years between 1983 and 2016 in construction, fell by a month or two in manufacturing, and has risen across most intervals in retail trade and finance. Tenure in government was up by nearly two years between 1983 and 2016. Of course 1983 was not 1963, the Golden Age before everything reportedly went haywire in the 1970s. But 1983 was very early in the neoliberal era, so we should expect to see more of a change than we have.


Another counter to the job-churning story: turnover now is lower now than it was in the early 1990s. The monthly employment numbers—like “employers added 223,000 jobs in May”—are net figures, and are the difference between many gross job gains and losses. At current rates, that 223,000 figure probably results from something like 2.5 million gross gains, from new and existing employers, and 2.3 million losses, from shrinking firms or ones that go under. (We won’t know the actual numbers for many months yet—these are just guesses from recent trends.) Add the gross gains and losses together and you get figures on gross turnover—almost 5 million jobs come and go in a typical month lately.

Here’s what those gross numbers look like (translated into percentages of employment to make them comparable over time). All are down considerably from the 1990s. The trendline in the turnover graph heads steadily down; the r2 figure of 0.92 means that the trend alone explains 92% of the movement in the series. That’s not to say time is causing the downtrend; it is to say it’s very well-established.

BED gains, losses, turnover

part-time work

Another claim: we’re all juggling multiple part-time jobs, or working short hours because full-time employment isn’t available. There has been little change in the share of part-time employment in the total since the early 1980s. It was around 17% of the total in 1980, and is around 17% today. Part-time employment comes in two measures: noneconomic or voluntary (people who prefer part-time work), and economic or involuntary (people who want full-time work but can’t find it). Noneconomic is essentially trendless since 1975. Economic is highly cyclical, rising in recessions, and falling in expansions. It’s now at or below the lows of the late 1970s and late 1980s. It’s slightly above where it was in 2000, a serious boom year. (There were definitional and technical changes in the survey on which these numbers are based in 1994; they took the economic share down by 1.1. Instead of inferring whether part-time workers wanted full-time, as had been the case earlier, surveyors began explicitly asking if respondents want full-time work, a not-unfair question but one that yields smaller numbers. There have been no significant changes in the survey since then.)

part time

multiple jobs

Ok, maybe gig work isn’t the norm, but lots of people are doing it on the side, right? Maybe, but you’d have to find some fresh numbers to support that case. Here’s the share of multiple jobholders. It declined from 1998 through 2011 and has been flat since. Lately less than 5% of the employed have been working more than one job. We’ve only got 24 years of this data, so it’s hard to make generalizations about it, but it looks like multiple jobholding rises early in an expansion and falls later as the labor market tightens.

Multiple jobholders

temp work

There has been no uptrend in temp employment in almost 20 years. There was a sustained rise from when the stats on temporary help firms (then called “help supply” in BLS argot) begin 1982 through 1998, but the trend has been flat since. Temp employment fell sharply in the Great Recession, and somewhat less in the early-2000s downturn, and rose after both to the 2% neighborhood. Temp firms accounted for 2% of employment in May 2018, exactly what they did in October 1999.

Temp employment


If a growing horde of us were 1099ers (disclosure alert: I’m one), then you might think the count of the self-employed would reflect it. But they’re not showing up in the BLS’s monthly employment numbers (which they compile by asking 60,000 people lots of questions every month). Self-employment has been declining as a share of the total for almost 70 years, with the exception of a mild rise in the 1970s and 1980s. Yes, this is also about people’s main job; there may be more freelancing to pick up extra cash. But these numbers offer no support for claims that the traditional job is disappearing.


As I said at the end of the first contingency post, none of this is to say that the world of work is pleasant or secure. Much of the population works hard and still can’t make ends meet. Except at the top, wages have been stagnant for decades and benefits are disappearing. But the stories of a new precarity are grossly overdone. You have to wonder that they gain salience because they’re written by academics and journalists, two fields where contingent employment seems more prominent than in other fields. But precarity has been the condition of the working class since the beginning of capitalism. That’s a major reason why it’s a bad system and should be replaced.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: