Can Olive Garden Bring a Divided Nation Together?

An increasingly central question for an increasingly divided nation is how to promote more interaction and understanding between socioeconomic classes that seem more and more out of touch. A new working paper from Maxim Massenkoff of the Naval Postgraduate School and Nathan Wilmers of the MIT Sloan School of Management offers an unexpected answer of sorts; spend more time at affordable, casual restaurants. The paper, called “Rubbing Shoulders: Class Segregation in Daily Activities,” finds that one of the few areas of common ground for high-income and low-income Americans are chain restaurants such as Olive Garden and Chili’s, not churches or post offices or pharmacies, as some might expect. Massenkoff spoke to Spotlight recently about the paper; the transcript of that conversation has been lightly edited for length and clarity.

Congratulations on the paper. Tell me how you got started on this—where’d the idea come from?

My co-author and I had been learning about this SafeGraph data, as there was a big rush by different technology firms to provide it in 2020 to help track lockdown compliance.

And it’s anonymized, correct?

Yes, you typically only see it at an aggregated level. Instead of saying this person was here at this time, it’s more like five people from this neighborhood go to this other neighborhood. It seems impossible to me to figure out someone’s specific identity and I think they’ve put in a lot of work to ensure that.

So, it was just amazing to me as a social scientist that you can use this data to look at where people go, something that we usually don’t have very good information on. It opened up a lot of different opportunities. And the first work I did with the data was thinking about it with respect to crime. During COVID, there was a big drop in per-capita crimes, but in a time where activity was so depressed, it seemed like that might give you a misleading view on public safety because if no one’s outside, then that will just kind of mechanically decrease the opportunities for crime.

So, we thought what if we use SafeGraph to calculate a new kind of crime rate that’s based on foot traffic? Because the reason we calculate crime per capita in the first place is that people want a measure of risk like that—that’s why we don’t just compare the total number of crimes in New York City to San Francisco. It doesn’t really make sense to do so because the cities have such different populations.

That was our first work with SafeGraph, but then we thought, there’s a lot of work in social science on, interactions across class lines and the importance of having connections that are different from yourself and how that might connect to economic mobility, for example. And because the SafeGraph data was tracking where everyone was from and where they were going, it seemed like a really cool opportunity to just ask a basic question, which is, where are you most likely to encounter somebody different? That was kind of the motivating question for me; if we really value these encounters across class lines, we could for the first time look at the data and say where those encounters are most likely to happen.

I was really motivated by this question of having more interactions across class lines; that’s part of what cities are supposed to facilitate. In the paper, we’re careful to not say that we’re certain that more of these interactions need to happen. But if you asked a bunch of social scientists to bet on whether or not this kind of mixing is good, I think, I think they would say that overall, it’s good.

Tell me more about what you found and how surprised you were or weren’t by that.

We found, that first off, the rich and poor are isolated. In their typical visits to an establishment or a park, they tend to encounter more people that are similar to them than you would’ve expected if everyone was randomly going to places. That’s far less true for the middle-income group. By the lights of our data, the wealthy are actually more isolated than people in the bottom quintile just in terms of who they encounter in their daily activities.

We then try to break that down and look at why that is—and a really obvious candidate is proximity. People tend to go to places that are close to home and the U.S. is very residentially segregated. There are neighborhoods where there’s a mansion on every block and so everyone going to the post office on that block is probably going to be really wealthy. Proximity was the biggest single factor that we identified. If we assume, for example, that people go to pharmacies the exact same amount, it actually wouldn’t change isolation that much because everyone chooses their local pharmacy go to. And that makes it highly likely that your pharmacy’s going to reflect your own income.

So, if you were looking for a policy solution, it’s much more driven by housing policy than anything else.

Exactly. If you look at what HUD has been putting out and the kind of things that they’re encouraging, there’s this idea of mixed-income communities. If you really want people to congregate under the same roof, then it would be a super attractive idea to have more mixed housing, though regulations can make that hard.

And for churches, I’m assuming you did not find that that was a place where different classes tended to mix at least not as much as you thought that you would?

Compared to the average visit, a church is more isolating for the people in our data set. The net effect of churches is to increase income segregation, and that was a surprise. Distance, again, is a big part of it. I was looking at a subsample of data and people tend to go to churches that are really close to their homes.

But the Olive Gardens of the world are an exception to the rule?

Yes. I talked about how the rich and poor are isolated, that we think a lot of it has to do with proximity and very little to do with industry. But people obviously go to different brands. And so, in the next step of the paper, we go a level deeper and try and unpack the role of brands. And so, this just comes from an exercise where we delete that restaurant or chain from the data set and just recalculate what isolation would look like for the different groups. So, let’s assume you were still going to every other place in the same proportions, but now Olive Garden shut down, would that lead to an increase or a decrease in integration? And one of the things that stood out was how these casual, full-service restaurant chains, have a positive effect on both ends of the income spectrum Places like Olive Garden and Chili’s help integrate the lower-income groups and they also help integrate the higher-income groups. This wasn’t something that we were expecting; we didn’t come into the analysis with any kind of Olive Garden hypothesis.

I’m sure.

Though when we were doing the research, we did take pleasure in the fact that Olive Garden used to use the slogan: “When you’re here, you’re family.”

In terms of policy implications, we really don’t want to take any next steps. We’re just trying to put these rankings out there and suggest that mixing might be one thing to consider as we’re thinking through different policies. But obviously, the fact that CVSs tend to isolate people wouldn’t mean that the solution would be to close down half of all the CVSs. The objective of paper is not to say mixing is everything.

In the last section in the paper, we look at the kinds of people who patronize different kinds of establishments. And here, we just really wanted to flag this for other people who might use the data because there’s a lot of discussion about the pace of gentrification and what the permitting process should be like for different kinds of businesses. With this new data, you can look at the kinds of people that go to different kinds of establishments. And one of the things that we report is that the average chain in the data is attracting more people from low-income and majority non-White neighborhoods compared to the average independent establishment. For permitting laws, if the goal is to slow gentrification, then it might be useful to check in on this data, because right now what it’s saying is that the typical independent store is bringing in more people from rich or high-income areas.