Y Combinator has been called the world’s most powerful startup incubator, but now it’s testing something completely different: investing in people’s lives, not their businesses. A few weeks ago, the startup incubator announced a pilot program on basic income in Oakland, California.
The idea of a basic income — a stipend guaranteed to each person to cover essential needs — is gaining momentum. Finland is considering it. So are four cities in the Netherlands. Canada is planning a provincial test project in an Ontario community. In Switzerland, voters recently rejected a proposal introducing basic income, but it’s notable that the vote happened. Many details of the Oakland project need to be worked out, but the premise is simple: A group of residents will be given stipends, and Y Combinator will research how that income affects the recipients.
Y Combinator’s ambitious project is unusual because it’s a private organization — not a government — initiating a study on a social-safety net. “We want to run a large, long-term study to answer a few key questions: how people’s happiness, well-being, and financial health are affected by basic income, as well as how people might spend their time. But before we do that, we’re going to start with a short-term pilot in Oakland. Our goal will be to prepare for the longer-term study by working on our methods — how to pay people, how to collect data, how to randomly choose a sample [participant], etc.,” a Y Combinator blog post says.
The U.S. hasn’t seen this much chatter about basic income in decades. Richard Nixon supported a version of basic income in 1969, and although he called it “guaranteed annual income” to make it sound less pinko, it died on the Congressional vine. Opponents of basic income tend to argue that the programs would be too expensive, and they would make people lazy. (MIT Technology Review recently called basic income “a sellout of the American Dream.”) Basic income supporters say it will alleviate poverty, simplify government assistance, and help people when robots jack their jobs.
Both sides wonder how the hell the Y Combinator project will work, which is why I spoke with Elizabeth Rhodes, the program’s research director, about how the company plans to navigate the complexities of the program.
Kate Knibbs: Does Y Combinator know how it will measure things like happiness and productivity?
Elizabeth Rhodes: We’re working on developing our outcome measures. So the pilot that was announced last week is really just a test. I mean, we’re not going to be able to answer any research questions with it, it’s going to be really short-term, and we’re just going to test out all the mechanisms — how we’re going to transfer payments, how we’re going to survey and measure everything. There are fairly standard measures for some of these survey questions. We’re working with a lot of different people who are experts in … particular areas that we’re interested in who have developed measurements. And especially for the longer study, we’re going to use as much administrative data as we can. Participants will be invited to sign a waiver … so we can access as much data that’s being collected [as] passively as we can without bothering people.
Would the participants allow you to see their employment status and how much money they were earning with their jobs? How detailed would your data collection get?
For timing or just in general?
In general, but also with metrics like productivity.
Well, productivity — we can’t have an outcome measure that’s just productivity, because it’s so subjective. We want to look at how people spend their time, for sure, whether they’re working. And we’ll be doing surveys on an ongoing basis to get measures, certainly at a point where we can get that data without bothering people, about things like how they spend their time, whether they’re volunteering in the community — if they have children, if they’re spending more time with their kids. In general, we’ll kind of ask them questions about what they’re doing with their time with specific areas that we are thinking about in terms of volunteering, community engagement, parenting, working, things like that.
So responses will be mostly self-reported? Y Combinator won’t be looking at their phones to see how much they’re calling their parents or anything like that?
No, no. Mostly self-reported. Some of it we can get, like employment hours, passively, through administrative data, if they’re willing to sign waivers so we can corroborate that, but a lot of it is asking them questions.
The project is not going to be income-based, but how is Y Combinator going to figure out whom to give money to? Will it be from a pool of people who have applied, or will the company randomly select members of the community?
We’re going to make it as random as possible. So we’re working on figuring out a sampling from which to draw from. The goal is not to have people apply, since that would introduce a lot of bias. … People would obviously be able to say no if we reach out, but we are going to try to get a random sample. I imagine for the actual study there will be some sort of parameters, you know. It’s not going to be income-based, but there might be a cap on income, or household income. … At some point, we’re going to have to limit variation if we’re going to have some kind of statistical power, but those are decisions we’re going to be making during the pilot, and just in general, we’re working through some of those questions now.
Would people who have a criminal record be barred from the program?
No, they would not be excluded.
The blog post about the project says that if the pilot program does well, Y Combinator plans to expand it. I’m curious what the criterion for “doing well” would be?
I think there’s a couple different things. One is, we want to, while talking to people during the pilot, see if the design we’re coming up with approximates basic income. Are people able to meet basic needs with the amount of money that we give? The goal is to make it as closely approximated to basic income, but also, because we’re not giving this to everyone in the city of Oakland, because of social pressures … if everyone were getting it, would things be different? They [would] have a lot of pressures from family or friends about loans. … Just to make sure it’s not harming people in some way or making things more difficult. And even, more generally, are we going to see results? With a five-year study, would we be able to draw any conclusions about basic income? And if our designs don’t seem to be working, we’ll revise — it doesn’t mean we won’t do a larger study, it just means that we will make tweaks accordingly.
Do you know how much money you’re going to give people?
We don’t yet. It’s likely to be between $1,000 and $2,000 a month. And where that falls, we’re working with a lot of different people to figure out. We recognize that [by] choosing to do it in Oakland, we’re going to have to give more money than we might have to in other places in order to meet basic needs, especially because of the cost of housing, so we’re considering that, and recognizing that it might be higher. So when people hear a number and think about how that would play out nationally, it might be different.
Were high housing costs part of the reason Y Combinator decided to do the pilot program in Oakland?
Partly. … The cost of living and housing is more expensive, so it makes the study more expensive. But at the same time, there’s a concentration of wealth and vast inequality, so in some ways, it’s a microcosm of what might be coming, and what we’re already seeing with income inequality nationally. And with technology and developments [continuing] to put people out of work and then there’s this growing wealth among those creating the technology, [more and more places could look] like Oakland. So I think it’s … perhaps not the most representative of the country at this exact moment, but it’s an interesting place to experiment with this.
Will the people who receive the money be tasked with reporting how they spent it, or will Y Combinator have direct access to that information?
Those are questions that we’re still thinking about and working through: what is OK to ask people and what is not. Those are research-design questions we’re working through right now.
How did Y Combinator decide on the timeframe and the amount of people for the pilot?
It’ll be less than 100 [participants], and we’re not sure how many exactly. But we wanted a large enough sample to be able to get some insight into research design and whether it “works,” but we don’t want to run it for too long. With [a timeframe of] six months to a year, people will be like, “You’re not going to be able to see anything. You can’t quit your job for six months to a year.” And we understand that. So [the initial pilot is] not about research questions, it’s more about the study design and how it’s going to work and our various measures. We didn’t want to run it too long because we wanted to get started with the larger study. … It depends on how things are going and whether we feel like, after six months, that we have a strong business design or there are things we need to be tweaking. I’m hoping to start fairly soon recognizing that we’re not going to have it all together. .. We’ll be constantly tweaking with the pilot as we go along, which is why that data won’t be used at all in the [full] study.
When you say tweaking, will the amount of money people are receiving potentially change?
It could, yeah. By the time the study starts — the full study — we’ll have it set. But we’ll be making changes if things are not working or not optimal. We’ll make changes until we get it right in the pilot.
What’s the company’s relationship with the city of Oakland? Is it working directly with the city council? Is it more of a casual relationship?
We’ve been talking to people. … It’s very casual at this point. We actually just decided to do it in Oakland like about a week or so before the blog post went up, so it’s very new. We’re starting to reach out. We definitely want to work with lots of community groups and individuals in the city, so we are beginning that process. We don’t have formal partnerships in place but we are reaching out.
Has Y Combinator has spoken to any members of the Oakland City Council?*
[*I’ve contacted all the members of the Oakland City Council and asked whether if they have spoken with Y Combinator; Councilman Dan Kolb said through a representative that he had not. “I have not spoken to them,” Vice Mayor and Councilwoman Desley Brooks said via email.]
If this pilot goes well and the project is expanded, are there other cities or locations that you’d be interested in running the next stages in?
We’re not sure, we’re going to see. if things go well in Oakland, [it] might just be expanded in the Bay Area. We considered a number of different places before deciding, so right now we’ll just see how it goes.
The initial study will include less than 100 people. Is there a floor for how low the number could be?
I’m not sure. … I’m sure we’ll have a group that receives an income in a control group for the actual study, but I don’t think we’d go under 30 people receiving an income and 30 not. The numbers are very up in the air at this point.
I know that some Y Combinator–backed startups have ties to the financial world, and I’m curious about how autonomous this project is. Will you be working with those startups?
No, we’re definitely on our own. … YC Research is the nonprofit that Y Combinator founded, and I work for YC Research and we’re doing the study through YC Research.
Are there any pilot projects about basic income or about economic security that you’re drawing inspiration from?
We’re definitely talking with GiveDirectly, [as well as] people who are working on the Finnish study, and the Netherlands, and the Canadian study. We’re definitely in conversations with other people who are working on studies all over the world. … We’ll definitely be cooperating and collaborating hopefully on outcome surveys, and data that is useful. But I think we’re finding our niche. Finland is much larger — I mean, the government is funding the study — and looking at employment and how [basic income] affects work, they’re not really measuring things like subjective well-being. So we’re all doing slightly different things, but in conversation, and trying to complement the work that we’re doing.
Would Y Combinator want a partnership with the California state government? The project is interesting because most other basic income projects are affiliated with governments. This one’s private. How could that change things — would Y Combinator turn down a partnership with the government because it is specifically interested in seeing how a private model would work?
I certainly don’t think we would turn anything down. … [This is] a first step. To really get a fuller picture you need to look at macro-level effects — if you gave it to everyone in the city, what would be the effect on the economy and jobs [be] and things like that. … You can’t do that with a smaller study. But right now, it’s just getting popular. I’m not sure governments are quite ready to take on a big study like this at this point. So we see this as a first step. If we find promising results or we learn things that could inform future research and policies, that would be great. It’s not something where we’re like, “We want to do this by ourselves.” It’s definitely more: Let’s get this started and see where it goes, so we can take on partnerships and encourage others to study it, as well.
Is the money coming directly from the Y Combinator research department’s budget or the company’s as a whole?
The pilot is being completely funded by the research department.
Did Y Combinator get grants for the project, or did is the funding coming through a pool of money?
It was started with money from Y Combinator. … I think at some point after the pilot we’ll be looking at fundraising or apply for grants. But right now, we’re completely self-funding.
This interview has been edited and condensed.