The Covid-19 pandemic has accelerated the tech-driven transformation of the grocery industry. Many Californians shunning brick-and-mortar stores for fear of catching the virus continue to rely on workers hired by online ordering apps to pick, pack, and deliver their groceries. Exposure to Covid-19 is far from the only risk these workers face.
California will be voting on Proposition 22, a ballot initiative that would roll back minimum wage and other labor protections for app-based and other gig workers. Their rights to a decent living and safe and healthy working conditions hang in the balance.
Late last year, California passed Assembly Bill 5, which requires many gig companies to reclassify their workers as employees instead of independent contractors. Under AB5, workers are entitled to the state or local hourly minimum wage and other protections that come with employee status, such as paid sick leave and unemployment insurance.
AB5 sets an important precedent for how to protect workers’ rights as work changes. But a handful of gig companies have launched a multimillion-dollar campaign urging Californians to vote for Prop 22, which would gut many of the law’s protections.
If this effort succeeds, our research shows workers in the online grocery industry could face financial hardship, hunger, and even destitution.
The lack of minimum wage protections for independent contractors has made it possible for the popular grocery shopping apps Instacart and Shipt to introduce opaque algorithms that trigger unpredictable reductions and fluctuations in shoppers’ earnings. These algorithms replaced transparent formulas that were also linked to higher pay.
In October 2018, Instacart launched a new pay algorithm to account for factors such as the number and type of items ordered, store location, and distance traveled. But the company has not disclosed the full list of factors that influence workers’ pay, or how each factor is weighted in the final calculation.
Some workers have raised suspicion that the algorithm may be identifying the lowest pay people are generally willing to accept in the region where they work. While an algorithmically generated “reservation wage” could pave the way for pay discrimination, Instacart’s secrecy makes this impossible to verify.
Instacart shoppers we spoke to described baffling reductions in their earnings after the new algorithm was introduced. A 35-year-old shopper from Los Angeles who asked that we not use his name told us that, in November 2018, his weekly earnings fell from $1,000 to $800, even though he was consistently working at least 60 hours a week. This meant that his earnings dropped to about $13 an hour, before accounting for his expenses.
Like all shoppers, he pays for work-related expenses out of pocket. Monthly expenses include gas (about $400), car insurance ($100), a phone plan that can cope with data-intensive apps ($150), and other essentials such as cooler bags and cleaning equipment to wipe down groceries.
After accounting for these costs, he made less than the minimum wage in Los Angeles, which was $13.25 per hour in 2018.
To cope with the loss of income, he tried cutting back on essential items on his own grocery list. “I guess I won’t buy eggs this month. I’ll just get half a gallon of milk.” He also struggled to make rent, and worried he would become homeless.
His experience is far from unique. In 2019, the labor rights organization Working Washington analyzed 1,400 weekly earnings reports from Instacart workers across the country, to work out how much they were earning before tips. After accounting for gas, business expenses, and additional payroll taxes borne by independent contractors, the group found that the average pay was only $7.66 an hour. Approximately half of the weekly earnings reports were below $7.25 per hour, the federal minimum wage.
Shipt, a subsidiary of Target and one of Instacart’s main competitors, also uses an algorithmic pay model in California, which has just been rolled out nationwide. Shipt’s algorithm is also largely a mystery, but the company claims that it generates an estimated pay range that reflects the “effort needed to shop and deliver an order,” considering factors such as estimated shop and travel time.