The Brain Gain: The Impact of Immigration on American Innovation

Kevin Cool: If the United States wants to remain a global hub of innovation, then we need to understand the role of immigrants.

Tracey Panek: Here’s our vault museum. This is where we keep all of Levi Strauss and Co.’s treasures that go back to the 1800s.

Kevin Cool: Meet Tracey Panek.

Tracey Panek: I’m the historian at Levi Strauss and Co. and director of our archives. Oh, we’re going to have to open it up and then do it again.

Kevin Cool: Tracey is standing in front of a blue fireproof safe. She’s one of only two people who know the combination.

Tracey Panek: Okay, you must avert your gaze. Inside we’ve got shelves and we have everything from our XX which is an overall from the 1870s to 9 rivet. That dates back to 1873, 1874. So, essentially what you’re looking at here are the oldest, the most important pieces in our collection.

This is where the designers come in. We’ll pull out pieces from our fireproof safe. We’ll bring out other items. And they’ll look at everything from the stitching to the back labels. They will look at anything that will help give the inspiration for a new piece or a portion of a new piece. Our designers really get their inspiration from our archives in order to create something for the future.

Kevin Cool: Immigration is a hot topic. It’s the foundation of the American experience, and it’s also a source of political division. But immigration’s role in American innovation is poorly understood, not just now but historically.

Tracey Panek: Immigration and the immigration story related to Levi is really such an important part of the Levi Strauss story. Here was a man who came to America, saw an opportunity to create something for himself where there hadn’t been opportunity in his home. It just wouldn’t have happened.

Thank goodness our rivets are still there. And with a magnifying glass you can see LS &Co. May 1873, the date that we took out our patent.

Kevin Cool: But Levi Strauss’ isn’t just about immigration and opportunity in America, it’s also a story of collaboration.

Tracey Panek: Levi was an immigrant from Bavaria. He arrived in America first in New York. It was an era when there were a lot of restrictions. He came from a Jewish family. He was the youngest son. There were few opportunities for him to either find work or to marry. He’d been a naturalized American citizen before coming out to San Francisco. In the 1870s he had customers all over the American West, a pretty solid network of retail shops.

And one of his customers was a tailor who lived in Reno, Nevada. Jacob Davis, another immigrant, he had come up with an unusual way of making workpants. And his idea — tiny but important — was to take a little piece of metal that he put into the pockets of a pair of pants so that when you put your hand in and out of them, they didn’t tear. That tiny but important innovation, or riveting pants, was the first time it had ever been done.

And he wrote a letter in 1873 to Levi Strauss & Co. And Levi, although he hadn’t done any manufacturing to this point — he’d been an importer and exporter — he agreed to take out a patent which is what Davis proposed. Levi invites Davis to move to San Francisco where the company is headquartered, and Davis will oversee the manufacture of the first riveted pants, or the 501s as they came to be known.

I like to think of Davis and Levi creating 501s as the company’s first collaboration. Now we do hype collaborations with big name brands. But that’s where it began, partners working together to create something special. Let’s take a look at 9 riveting.

Kevin Cool: It may seem obvious that immigrants make enormous contributions to American innovation. What’s less obvious is that they are also valuable collaborators who make their colleagues more productive as well. Levi Strauss and Jacob Davis exemplified this more than 100 years ago, but is it still true today, and can we prove it?

This is If/Then, a podcast from Stanford Graduate School of Business, where we examine research findings that can help us navigate the complex issues facing us in business, leadership, and society. I’m Kevin Cool, Senior Editor at Stanford Graduate School of Business. Today we speak with economist Rebecca Diamond.

Rebecca Diamond: I’m the class of 1988 professor of economics at Stanford Graduate School of Business. I research labor, urban economics, and adjacent topics such as innovation and immigration.

Kevin Cool: Our if/then statement for this episode is if the United States wants to remain a global hub of innovation, then we need to understand the role of immigrants. What do you think is poorly understood about immigrants and innovation?

Rebecca Diamond: It’s been very hard to quantify how large of a role immigrants play in innovation created in the U.S. There’s a lot of anecdotal evidence or small data points that indicate they play an outsized role, such as the share of Nobel Prize winners that live in the U.S. are disproportionately immigrants. But that’s obviously a very small sample of people to study. Innovation in itself is just an elusive thing to measure.

So, we take the more modest goal of looking at what is the role of immigrants in producing patents, so that’s at least a quantifiable thing. And we don’t really know the role of immigrants in patents because the data that tracks patents doesn’t tell you much about the inventors that are part of those patents. It tells you their names, but you don’t know who’s an immigrant and who’s not. So, we found a new data source and a new way to link data together to actually measure that.

Kevin Cool: So, patents as a measure of innovation, and we’re talking about highly skilled immigrants. First of all, what do we mean by highly skilled immigrants? And then if you can unpack the data that you used and how you validated who was an immigrant in that inventor cohort.

Rebecca Diamond: So, we don’t have any direct measures of skill. We just have people that are inventors of patents. So, we’re sort of just looking at inventors, which we think of as high skill people. And then to be able to quantify who is an immigrant in our data, we start with the standard database that many people have studied which is the patent database.

So, you have for every patent in the U.S. going back to the ’70s, what the patent is, the names of the inventors, some details about the technologies that’s been invented. And the only thing you really know about the inventors is their names and their city and state of residence. I don’t have a patent, but if I did, Rebecca Diamond, Menlo Park, California, would be all you would know about me.

And so, the way we’re able to get a better angle on the inventors is to link it to this database of migration histories is what it’s mostly used for. I’ve used in previous work studying. So, what the Infutor data is, is it’s your first and last name and then the addresses of where you’ve lived over time and move dates. When you link these together, we’re able to link a name and city and state of residence at the time of the patent being applied for.

And then the next step is to figure out, well, who are these inventors. That doesn’t tell us who’s an immigrant and who’s not. That just says I can now track your patenting behavior from age 25 to age 65. And there is this additional piece of information that’s also in this data. So, the data also contains the Infutor data the first five digits of one’s Social Security number. We have it for about 80 percent of the population. You can figure out from those first five digits the year that those numbers were issued.

Kevin Cool: So, it’s like a timestamp.

Rebecca Diamond: It’s a timestamp that’s embedded in the number. So, if you know the formula, you can back out when this number was issued. And that tells you the year that you got your Social Security number. So, how is that useful for us? That’s useful because if you’re a U.S. born American, you either get your Social Security number when you’re born, which is, I think, true basically since the 1980s. I got my SSN at birth.

If you’re older than me, you got your SSN at the time you first were employed. And luckily, historically teenagers worked at a pretty high rate — not full time, but they had summer jobs or odd jobs — less true today. So, almost everyone is getting their SSN before they’re 16, historically. So, then we can identify the people that get their SSN at age, say, 20 or later. So, if you’re getting your SSN assigned to you at age 20, highly likely that you’re an immigrant.

Kevin Cool: Okay.

Rebecca Diamond: So, we define immigrants in our sample as people who get their SSN at age 20 or later.

Kevin Cool: Just to be clear then, this is publicly available data and you’re only using the first five digits of the Social Security numbers in the study. So, what was the evidence that suggested that immigrants play a disproportionate role in patent activity?

Rebecca Diamond: Once we’ve built this database, we’ve built it for the whole Infutor sample, not just for the inventors. We can do this for anyone in that data. So, in that data, I can first tell you what share of the U.S. population covered by this data is an immigrant regardless of whether you’re an inventor. And we have that number; I think it’s around 10 or 11 percent. And then we can look at conditional on issuing at least one patent, what share of those people are immigrants. And that number is higher. I think it’s like 15 percent.

Kevin Cool: 16 percent I think is what —

Rebecca Diamond: 16 percent. So, that’s already telling you that if it was just the same as the typical U.S. population, then the share of inventors that are immigrants should be the same share as Americans that are immigrants, but you see that the share goes up when you condition on being an inventor. And then we can also look at how many patents you get. And we can find that even conditional on being an inventor, immigrant inventors have more patents over their life cycle than Americans. Each margin where you raise the bar, you find an increasingly disproportionate role of immigrants.

Kevin Cool: And the figure in your study was 36 percent of the innovative output based on patents between the years that you were studying, which was 1990 and 2016, were ascribed to immigrants.

Rebecca Diamond: If you just look at what share of patents are done by immigrants, I think it’s about 24 percent. The 36 percent comes from when you account for those indirect effects of immigrants as well.

Kevin Cool: You’re listening to If/Then, a podcast from Stanford Graduate School of Business. We’ll continue our conversation after the break.

Can you explain this spillover effect, that the innovation of inventors contributes to higher levels of innovation overall?

Rebecca Diamond: So, the sort of simple part of the paper — I mean, it was not simple simple, but once we just linked the datasets together, I could just give you some raw statistics like what share of patents have an immigrant author. Those are some of the stats we talked about. But the innovation production function is a lot more complicated, right?

The people that choose to collaborate together are choosing to, and there’s lots of unobserved things about those people and about their productivity that could lead to that collaboration. So, to be able to disentangle how much of this collaboration affects your productivity is very challenging because it could be the productivity affecting the collaboration or the collaboration affecting the productivity.

The gold standard way of estimating the causal effect of some treatment on an outcome is using a randomized controlled trial. This is very challenging in this case that you’d have to randomize people together, force them into a room and say, “Figure out how to write a patent together.” And then once they do, say, “All right. Now you’re free to go live about your life,” and see how their future productivity and innovation is changed from that forced collaboration.

So, it’s hard to think that you could ever get ethical approval to run that collaboration, actually implement it or have the funds to do something at scale, and it would take a long time. So, instead what we use is what you might call a natural experiment, something that’s naturally occurring in the world that creates something like a randomized experiment for us.

So, instead of forcing people into collaborations and looking at how that affects their productivity, we look at a setting where people are basically forced to no longer collaborate, despite having collaborated in the past, and see how that affects their productivity. So, what we look at is early inventor deaths. You and I have written a patent in the past. And if I am unlucky and pass away young, I can then see how are your patenting behaviors changed relative to someone who looks very similar to you and had collaborated with someone very similar to me, but their collaborator didn’t pass away.

Kevin Cool: So, there’s an attribution effect, in essence, from that inventor who passed away young.

Rebecca Diamond: Exactly, because we’re sort of forcing that collaborator out of your life. So, the death we’re thinking of as basically random with respect to the future productivity of that person’s collaborators.

Kevin Cool: So, what did you find? What was the outcome there?

Rebecca Diamond: So, we’re tracking a number of different patenting outcomes, but a very simple one is we look at the number of patents per year produced by the collaborator before versus after the focal inventor passes away, and then comparing that patenting activity to someone who looks observationally equivalent to you but doesn’t have their collaborator pass away. And we see that the collaborator of the dying inventor has a very similar patenting productivity pattern before the death when you compare the control collaborator who doesn’t experience the death.

And then when your collaborator passes away, we see this widening divergence of productivity where the collaborator who lost their partner who passed away, they decline in patenting output after the death relative to this person who doesn’t experience the death. So, that’s true in both regardless of whether the inventor who passes away is an immigrant or not.

But then we split the sample and look separately at immigrant deaths versus native born inventor deaths. Both lead to negative productivity consequences of their collaborators, but the effect is much stronger when your immigrant collaborator passes away. So, that suggests that the access and the knowledge of the immigrants is particularly valuable and useful for their collaborators.

And that’s consistent with some of the descriptive facts that we can see in the raw data. The patents that are authored by immigrants are more likely to cite foreign patents, so they’re more aware of knowledge produced abroad. Their patents are more likely to also be cited by foreign patents abroad. So, there’s this sort of global knowledge network that is more easily diffused into the U.S. patents though the immigrants.

Kevin Cool: So, how do we use this information? How does understanding this, what effect does it have? Have you had any reactions to this paper? What does it mean for policymakers?

Rebecca Diamond: I would say no one is that surprised that immigrants play a dipropionate role in innovation. I think people have thought that was likely true for a long time through the Nobel Prize facts. And we know that immigrants were disproportionately working in STEM fields; we knew that already. So, I don’t think the fact that immigrants play a large role in patenting shocks anyone.

The way to have successful innovation is not to just put smart people in a room by themselves and tell them, “Think hard.” It’s the collaborate and work together and create new ideas through the synergies of their knowledge. And immigrant knowledge seems particularly impactful on the U.S. born workers and inventors, and I think we potentially didn’t know how big of a deal that was, right? So, we’re finding 24 percent of innovation in terms of who authored the patents, 24 percent of that is authored by immigrants.

But when you account for the fact that they have these externalities on the future productivity of their U.S. born collaborators, and you do the decomposition through that metric, then you find that 36 percent of all innovation can be attributed to immigrants because some of the U.S. born production is attributable to the collaborations of their immigrant coauthors. So, that’s a big number, suggesting that any policies that would limit or lower the number of immigrants coming to the U.S. for these super high skill innovative jobs would have a large effect on future innovation.

Kevin Cool: Your research in the past has dealt a lot with economic inequality, housing. You had a very influential paper on rent control using San Francisco as the test market. Did any of that previous research lead you somehow to this particular study?

Rebecca Diamond: Research always has funny paths of how you end up where you are in terms of what you end up writing a paper on. We ended up here initially thinking we were going to write a paper on the migration location [decisions] of inventors and potentially the role of where you live within the U.S. impacts your innovation. That was our initial idea.

And the Infutor data tracks migration well. And we realized that we could use this clever trick with Social Security numbers to identify immigrants while we started to do that. And then that just led to, well, if we can figure out who the immigrants are in the patent database, we should write that paper first because that’s even more first order.

I came into this from a migration point of view. And then my immigrant coauthor, Shai Bernstein, who’s worked on innovation much more than I have, that before we get into migration, let’s actually just write a paper on immigrants. So, that’s how we wrote this paper.

Kevin Cool: Rebecca, did you always want to be an economist?

Rebecca Diamond: I went to college planning to do physics, and I have a physics undergraduate degree. I always enjoyed quantitative methods that could say something about how the world worked. So, as a high school student, physics seemed the most compelling to me.

My high school didn’t have an economics class. But I took some econ in college partially because you were required to take some social studies, and that looked like the most quantitative social studies. And I’m afraid when you have to read a lot of books; it’s not my skillset. So, I’d rather take something with math than not with math.

And the more econ I took, the more connected to the questions I felt that we were using math to answer. And the more physics I took — while I think physics is super interesting and important — I felt just sort of less connection with my quantum mechanics class and my statistical mechanics class, and that led me to want to pursue economics.

Kevin Cool: What do you find satisfying about applying your research inclination and your skills to solving the problems that you’re addressing here?

Rebecca Diamond: I really enjoy working with new data because you don’t know what it’s going to tell you or where it’s going to lead you. So, finding new datasets or linking datasets together that haven’t been done in the past has been one of the ways I’ve made progress, and I find that very fun and exciting.

I also tend to want to have some sense of a theoretical framework to interpret the data through. So, I would say I am by no means a theorist — I’m not an economic theorist: I’m an empirical researcher — but I tend to use a theory lens to interpret the data. And I find that combination leads to analysis and outcomes that I find the most compelling because you have a better sense about what the data are telling you.

Kevin Cool: Is that unexpected aspect of it what excites you about it? What’s the thing that makes you most excited about this research?

Rebecca Diamond: I mean, doing research, 90 percent of your days feel like failure. And if you can get 10 percent where you are excited and you actually learn something, you have to live for those 10 percent. Even those few days where like, oh, maybe we could actually figure out who’s an immigrant in this dataset and have it make some sense. The first day we actually got those results in, it was like, wow, that’s amazing that this actually works and no one’s done this before. Most of the days you just feel like you have no clue what’s going on, so you have to live for those few days where it really works.

Kevin Cool: But you got the 10 percent in this one, it sounds like.

Rebecca Diamond: Yeah. I mean, 10 percent is if you’re doing — that’s like you’re hitting above — that’s good if you can get 10 percent. Some papers I feel like it’s 2 percent.

Kevin Cool: If/Then is produced by Jesse Baker and Eric Nuzum of Magnificent Noise for Stanford Graduate School of Business. Our show is produced by Jim Colgan and Julia Natt. Mixing and sound design by Kristin Mueller. From Stanford GSB, Jenny Luna, Sorel Husbands Denholtz and Elizabeth Wyleczuk-Stern.

If you enjoyed this conversation, we’d appreciate you sharing this with others who might be interested and hope you’ll try some of the other episodes in this series. For more on our professors and their research, or to discover more podcasts coming out of Stanford GSB, visit our website at gsb.stanford.edu. Find more on our YouTube channel. You can follow us on social media at StanfordGSB. I’m Kevin Cool.