Can Gig Work Help Workforce Recovery? A Pilot Review (Part 1)

Summary: This post reviews data from a pilot program that taught people how to use freelance work to increase their monthly earnings. In month 6 after training, the median graduate had earned $2,375 in extra income. By month 9, gig work was about 20% of their total income. The initial success of the program is a proof-of-concept for how cities can use job training to avoid the more concerning aspects of the “gig economy” while also aiding in regional workforce recovery from COVID-19. This post is part of a series on the Future of Work from Tech4America.

Much of the debate about freelance “gig” work is often hypothetical: large tech platforms, such as Instacart and Upwork, contend that more flexible work would improve people’s economic security, while labor critics worry that the growth in independent contracting will result in lower pay and lack of benefits.

Unable to reconcile whether the growth of gig work is good or bad, some states, notably California, are proposing or have already severely restricted the availability of independent contracting.

Yet, in the middle of a global pandemic, increasing overall rates of self-employment can help cities bounce back from economic recessions. Some European nations, such as Germany, encourage unemployed workers to try their hand at self-employment. “Programmes aimed at turning the unemployed into entrepreneurs may be among the most promising for ALMP [active labor market policies], both in Germany and elsewhere,” concluded a study in the Oxford Bulletin of Economics and Statistics.

Could active labor policies that encourage gig work help American cities bounce back from the pandemic and meet their workforce development goals?

Fortunately, there was already a natural experiment to test this theory, where a few city workforce agencies, mostly in San Francisco and New York, trained workers in how to earn income through online platforms.

This report series explores longitudinal data collected on those participants and analyzes the potential policy impacts.

That is, interestingly enough, at the same time as California lawmakers were constructing a law to severely restrict gig economy work, some city agencies were engaging in effectively the opposite strategy of training residents how to be freelancers.

“We created a set of tools to help people leverage the gig economy to advance their career path,” said former director of the San Francisco Office of Economic and Workforce Development, Todd Rufo. “Freelancers are businesses, and our mission is to help small businesses as well as independent workers achieve economic success.”

Why This Matters

The troubling aspects of the gig economy are really about a smaller set of struggling workers. Polling of gig workers consistently finds that the vast majority prefer being independent contractors. One early survey found that the vast majority (around 80%) work part-time and about half work either 15 hours or less — they’ll occasionally pick up work to pay for an unintended bill, or save up for a discretionary expense (like a vacation).

Though the percentages are small, there are thousands of people who often find themselves forced into gig work because of severe economic distress, only to discover that income can swing wildly and there are no employee benefits. These are the very folks that often seek help from city workforce agencies.

The question is: can gig training help the very people who need help from the government?

The Top-Line Results And A Case Study

When looking at the results, it is important to keep in mind that the vast majority of job training programs do not work and workforce policy often only boosts employment a few percentage points.

• 60 (35%) said they had gotten a job in gig work (Of 169 participants who answered the survey round after the first month).
• After the first month, the distribution of self-reported income was $94, $260, and $500 for the 25th, 50th (median), and 75th percentile of earners. In month 6, the total income reported was $663, $2,375, and $5,500, respectively.
• Gig work was largely used to supplement earnings from employee jobs with part-time income. In month 9, the median earnings from gigs was ~20% of income for those working 15–34 hours.
• This is comparable to a previous report on trained gig workers, who were estimated to earn around $22,000 (or $1,800/month).

So, what does this look like in practice?

SamaSchool, the former nonprofit school that trained gig workers, highlighted one case study of a New York resident who turned to freelancing after graduating from an IT school:

Kristopher managed to take what he was learning through Samaschool and combine that with his love of fitness and computers to find work. He found contracts through two online work platforms: Field Nation and WorkMarket. He attributes the experience he gained through those initial gigs to getting his first major contract as a replay technical analyst with the NBA. The day after he signed his contract, he flew to California to start working. Shortly thereafter, Kristopher was hired by the NBA full-time. Kristopher continues to perform independent work once or twice a week. He attributes his success to the experiences he received from both Per Scholas and Samaschool. When asked what advice he has for other independent workers, he says, “I definitely encourage people to do more independent work, instead of waiting for ‘that job’ to come. For the NBA, I got second and third call backs because I had that early experience [with Field Nation and WorkMarket].”

There are some important qualifiers about the data:

First, this pilot is observational. There will not be a control group, so we don’t know whether these recipients were better off had they never done the program, nor do we know whether they wouldn’t have naturally found themselves doing gig work if it never existed.

Second, survey data is prone to bias and often unreliable. If negative response bias is the trigger, the survey would over represent people who were struggling and wanted to share their anger about the program through the survey; if there’s a positive response bias, it would mean that folks are more likely to fill out the survey if they are doing well. This is to say we don’t know whether the actual effect of the program is higher or lower.

Third, there was variability in the project. Some learning was online, while some participants only did in-person training at centers with computers and coaches. The total training lasted about eight hours and could include additional coaching, but was entirely voluntary. Would more intensive training make a difference?

Initial Conclusions

The data does show that active labor market policy, in the form of training workers for the gig economy, can increase income. As a proof-of-concept, some meaningful percent of struggling workers can learn how to augment their income through gig work.

And, struggling workers who are trained in the gig economy appear to earn a meaningful amount of extra income

Future posts in this series will explore the data in more depth, partly based on the reaction to this initial analysis. For previous posts on this series, which explores the gig economy and the future of work, follow this link.

Data Availability And Preference For Transparency

I’m a big fan of transparency and will happily share the data and code. If you’re interested in the code/data, email research [at] greg ferenstein [dot] com. The data is in a Google Colab Jupyter Python notebook. I will create a public instance of the notebook with you so you can analyze the results. The reason I’m doing it this way is that I’m interested in both how researchers might analyze the data and their actual results, so I prefer it to be in a way that is easy to share with others.

*My statement of conflict-of-interest is here.