An Uber driver making $3 an hour. A Homejoy housecleaner admitting to living in a shelter. A Fiverr advertisement seeking workers whose drug of choice is “sleep deprivation.” Stories of the misery of the American gig economy abound.

It turns out the problem might be global. In a three-year investigation, researchers at the University of Oxford and the University of Pretoria spoke with scores of online gig workers from Kenya, Nigeria, South Africa, Vietnam, Malaysia, and the Philippines. These workers use digital platforms to bid on and perform tasks such as transcription, photo editing, and data entry. Much like their counterparts in the United States, the workers reported liking the freedom and flexibility that the work offered. Some described it as lucrative, as well. “For me, it’s a high paying job,” Angel, a contingent worker from the Philippines, said. “I was able to afford an apartment, pay my own bills, my own internet connection, my own cable, paying for our own food, for my kids’ milk. That’s all on my own.” (The authors did not include participants’ last names.)

But like their counterparts in the United States, they also described their work as unstable, uncertain, light on benefits, isolating, and taxing. Nearly half believed themselves to be easily replaceable. They complained about racism and discrimination. Moses, a translator, described making it seem like he was based in Australia, not Nairobi, in order to woo clients: “You have to create a certain identity that is not you. If you want to survive online, you have to do that.”

Then, there is the problem of massive labor oversupply. Unlike for Uber or TaskRabbit, which operate in a given city with a constrained supply of workers, the pool of labor for such digital work is for all intents and purposes infinite. One contingent-work platform reported having nine times as many workers as necessary. A Filipino virtual assistant described the inevitable result: “I first set [my hourly rate] at $8, because that’s what my previous client was paying me,” the assistant told the researchers. “But I found it quite difficult to find jobs. So I set it at $4. And I think I even set it at $3.50 currently. So, I mean, if you don’t get a lot of invitations, you don’t have any other choice but to lower down your expectations, I guess.”

Still, the researchers said that deflated wages, disempowered workers, and scant labor protections were not inevitable features of the online gig economy. Digital workers could organize, or could own and control their own labor platforms, for instance. Certification schemes and stronger regulatory protections might also help.

I spoke with Mark Graham, a professor of internet geography at the Oxford Internet Institute, about his and his colleagues’ research into this burgeoning world of work—and how to create policies to protect such workers, too. The interview has been condensed and edited for clarity.

Annie Lowrey: Could you tell me little bit about how you found these workers?

Mark Graham: First of all, we picked six countries that we wanted to study—three in southeast Asia, three in sub-Saharan Africa. We wanted to find workers who lived in these countries who were doing work on online gig platforms. Basically, we went on the platforms and created jobs. The job was: “We’re researchers who are studying this topic, can we talk to you?” That was our foot in the door. Then, some people we spoke to said, “You should speak to my cousin. He also does this work.”

Lowrey: In the paper you mention that this work is growing rapidly right now—something like 25 percent per year. How long has it been available?

Graham: There have certainly been platforms that have allowed people to work from anywhere, for anywhere, for some time. I remember there was a site, I think it still exists,, that allowed people to outsource little programming tasks.

What’s interesting is that now, we have so many people getting online at the moment and most of that growth in access is coming from the developing word. A lot of folks are coming online from low- and middle-income countries. While the idea of a website where you could outsource a gig isn’t that new, it’s that combined with the fact that we’re getting so many people who are willing to do lower-wage tasks coming online. That’s fueling some of the massive growth we’re seeing.

Lowrey: In some cases, it seems like a great deal for both sides, right? You’re in a high-income country with high labor costs and you want to get some transcription done. You’re hiring a worker from a low-income country where the wage seems competitive, even high. But you guys found a kind of wage deflation that people experience, due to the ever-increasing size of the labor pool.

Graham: That’s one of the most interesting things happening here. For a lot of the people who are engaged in it, this is good work. They have a job, they’re getting income. That’s great. For the client, who is outsourcing the work, that’s great too, because they’re getting cheaper labor costs and they don’t have the risks and responsibilities that they have with a more traditional workforce.

The issue arises when you think about this as a transformation that might be occurring in the whole economy. If we see this as a trend, if we see this as something that’s becoming more widespread, then there are problems with it, right? If you’re a worker in somewhere like the United States or western Europe, all of a sudden, your job is more at risk of being outsourced. The workers in places like Kenya or the Philippines or India, they’re getting work. But if we start to rely on this as a system, there are very few safeguards in place for people that rely on this as a source of income.

What happens if you become pregnant? What happens if you’re sick? Do we really want a lot of work that is so unregulated? It’s not looking out for the most vulnerable in any society. It’s good for those who thrive in it, basically, but it’s a problem for anyone else.

Lowrey: Did you get a sense of who is benefiting most from this system, and who is most vulnerable or benefiting least?

Graham: It is hard for us to say who is befitting most and who is benefiting least. But I think some of the people who benefit most are those with good levels of education, those with good language skills, and of course those who are in good health, which is key to all of this. Those who benefit least are those with poor education, poor language skills, and sadly, if they’re in poor health.

To go back to what you said, “It’s good for both parties.” It’s good for both parties when they’re thriving. If you ask a worker who has work or has access to work and is able to do that work, “Is this a good thing?" They’re probably going to say, “Yeah, it’s great. I love it.” But there’s selection bias there. If you spoke with someone who became ill and had to go to a hospital and then couldn’t deliver on time and got a bad rating, they’re going to tell you a very different story. We need to be careful to avoid that sort of selection bias.

Lowrey: Often, it seems that the worker is not getting the labor protections of the country where the work is being purchased or the country that they’re living in. They’re sort of like the Uber drivers who aren’t covered by minimum-wage laws. Is there really just nothing to protect these folks from some kind of exploitative, unstable, or even dangerous labor relationship?  

Graham: I think it’s worse than Uber drivers in the United States. Because with Uber drivers in the United States, at least you can make the argument that the work is supposed to be regulated, right? If you’re an Uber driver in New York, you could point to various rules and regulations that might apply. If you are an online worker in Kenya and the client who is giving you the work is based in the United States, it’s not fully clear to both parties whose rules should be governing that relationship. Should you be following Kenyan labor laws?

More needs to be done to enforce rules that are on the books in the first place. Then, more needs to be done to figure out how we make rules and guidelines and regulations make sense in a world of digital work. That’s obviously a difficult task that relies on regulators from all over the world asking, “How do we both encourage work but protect our workers?” That’s a hard balancing act.

Lowrey: Are there models to look to? Obviously, if I were to open a factory in Nigeria, I would be bound by trade agreements and by Nigerian law. But because these jobs are digital, global, unprecedented, do we have models to rely on

Graham: Step one is figuring out where the work is done. If you open up a factory in Nigeria, then most people would say, “That work is being done in Nigeria.” But if you have a business that’s in the United States and you’re hiring people from all over the place to log in to your servers located in the United States and do work for you in that manner, where is that work actually happening? I’d probably argue that work is happening where the workers physically are, where they’re living and breathing and tapping on a computer. But I think that’s the first step is to have a really clear sense of where it’s happening and make sure that any protections in place are actually applied.

That said, I worry that if we rely too much on laws that are already on the books, we don’t want to just encourage some sort of race to the bottom that drives work to the least-regulated places on earth. Because that also could happen.

Lowrey: You also describe a race-to-the-bottom on wages: Workers in middle-income countries feel like they’re being undercut by workers in low-income countries. And the short-term, spot nature of these contracts pushes that process along.

Graham: It’s a very common concern among workers. They have this very sharp sense that if they don’t do a piece of work, accepting the conditions offered to them, it can be very easily outsourced elsewhere. Is there any way around that at the moment? Not in the configuration in which all of this work is done now. But I don’t think that means that there are no possibilities to change that.

Lowrey: One of those things would be for workers to come together and exchange information or bargain collectively. You point to some examples of workers doing that—a Nigerian Facebook page, for instance. But this is transnational work, with a very dispersed workforce. If you go to work at a factory, everybody else who works at the factory is there and can see you agitating, or striking. But if the work is invisible, and all over the place ...

Graham: There are concentrations of workers in particular places. There’s a huge number of workers in the Philippines, for instance, especially in Manila. So if all of the online transcribers in the Philippines went on strike, it would still be hard for all of their clients to switch all of that work to a different place at any given moment.

That said, it is going to be very difficult for workers who have never met each other, who live on different sides of the planet, to try and collaborate rather than compete with each other. They’re probably going to undercut each other, they probably are going to take jobs from one another, in a world where there’s far more supply of work than demand for work. That’s why we need something along the lines of a fair-trade foundation for work, so that we can have more responsible clients who aren’t just trying to pay workers the lowest amount possible. It’s why we need sensible regulation. It’s why we can look to the platform cooperative movement, where there are corporations owned by workers, and looking out for the best interests of workers.