CCI Faculty Director, Karen Chapple, on the pandemic's effects on housing insecurity and displacement
CCI Faculty Director, Karen Chapple, on the pandemic's effects on housing insecurity and displacement
Professor Karen Chapple is the Faculty Director of IGS’s Center for Community Innovation (CCI). Currently, she and her team are conducting research on a wide variety of topics ranging from gentrification and displacement to local development under climate change. In July, IGS’s Kelly Jones had the opportunity to speak with Professor Chapple about two new projects centered around impacts of the COVID-19 pandemic on housing insecurity and displacement.
IGS: You are in the beginning stages of two new projects related to COVID-19 and housing. I know things are still tentative, but might you be able to share a bit about what questions are shaping these projects? More specifically, how do you view the relationship between evictions and displacement and the spread of the coronavirus?
KC: Sure. One project that we are working on, funded by C3.ai, is looking at what we call the precarity of the housing market in the US. We're developing what we call the Housing Precarity Risk model. What we're trying to understand, based on economic and health factors, is who is most at risk of losing their homes during the pandemic. It could be in the form of evictions, or it could be in the form of foreclosures, or actually, more likely, it will be a kind of invisible displacement. We always say evictions are just the tip of the iceberg. Probably for every one eviction that happens, there are 99 people that have been harassed and forced to leave in some way. And it can be really subtle, very simple. It can be a landlord saying, “I think I'm going to raise your rent in six months.” And even though you don't know they're going to do that, you start looking for another place. That's displacement and that's housing precarity. People don't have security of tenure. This is what we're seeing in the wake of this crisis: housing insecurity, which has been an issue because of growing income inequality nationally, now threatens to become a tsunami of change with people not being able to pay the rent because of losing their work.
IGS: Your project with the C3.ai Transformation Institute specifically focuses on using AI to track these sorts of trends in housing precarity. I am wondering if you could say a little bit more about how you are training the algorithms—how you’re using AI to track displacement and how your findings might be utilized?
KC: Yes, so we have three different kinds of big data or AI, or cloud computing, components of this project. One is actually being done by our partner, The Evictions Lab at Princeton, which is headed by Matthew Desmond. He's, the co-PI on the grant. Desmond’s team is scraping in real time the eviction records from a dozen cities across the US. The idea is that they will be able to put in place an early warning system, because they'll be able to see an eviction within just a couple days after the eviction notices were filed. They can then help policy-makers step in to avoid evictions. It could also serve as an indicator—a red flag—indicating other evictions that might take place in the same neighborhood. Desmond has collected eviction data for a long time. It's just that this is the first time he's done it in real time.
There are two other components of the project, which are AI intensive. In the Housing Security Risk model we're using mobility data to track people's movements. There is something that they call the “data lake” at C3.ai. (I love that name, “data lake.”) They've gathered all of the different forms of data that our researchers are using relative to COVID, and one of those forms is mobility data. We can look at activity patterns and we can find out, for instance, where there were commutes that are no longer. Then we can identify the homes of those people that are no longer commuting and figure out if these communities are at risk. That's one type of data that we can mobilize to identify vulnerable areas.
KC: Another thing we're doing in terms of AI is building a huge historical database. We're looking at mobility patterns over time. What neighborhoods do people—low-income renters in particular—tend to leave disproportionately (more than the average renter)? From where are low-income renters leaving or being displaced? And then we're looking at that over time and really paying attention to a very specific period: The Great Recession. This is an historical analysis. We're looking at 2008 and 2009 and figuring out what were the characteristics of those renters that were displaced in that time period or that lost their home to foreclosure? And then we can use that. We’ll update it a bit based on how neighborhoods have changed and other factors. Then we can train our data on that particular crisis in order to predict what the areas are going to be today.
IGS: Wow. Super innovative. You’re solving this problem by triangulating using a multi-pronged, data-driven approach.
Housing precarity is one of the most fundamental effects of this pandemic. Could you talk a bit about which populations have been the most impacted? For example, the IGS Poll’s last survey showed that disproportionate numbers of African American and Latinx California voters have been affected economically—but not only economically. They worry more about the effects of the pandemic in every domain of life. Are you seeing trends on the housing front that substantiate these worries?
KC: If you look at the preliminary work that's been done mapping areas that have been hardest hit by the pandemic, certainly they are areas with concentrations of communities of color. Most recently, what is coming out—and I've reviewed a few articles on this—what seems to be happening is, particularly in Latino neighborhoods where you have overcrowding, you're seeing disproportionate outbreaks of COVID-19. What we're learning is that, it's not about density. It's not about cities. You can have perfectly dense cities that have very low infection rates. It is about what's happening inside the home where people are crowded together. It’s also about crowding in other areas too, such as subways. And now we just learned that the baseball field is dangerous, too. We now have a whole team that's infected in Philadelphia. But crowding in the home is a really key factor. And one of the things that we know is that, particularly in the Bay Area, Latino families are crowded into very small quarters. We have been studying displacement for 10 years and one of the patterns we noticed early on—it's always held up—is that yes, you have Latino families displaced from places like the Mission. But they tend to reappear within San Francisco. They’re displaced to places that are just minutes from their home. How are they doing that? It's because when you lose your home, you move in with mom or you move in with your cousin, and your cousin lives on the next block. That’s how this crowding is taking place and a lot of the census data doesn't really tell that story. We really have to go to new forms of data to track crowding.
IGS: What impact do you hope your research findings will have on policymaking at the local level?
KC: In that Eviction Lab piece, highlighting evictions in real time is a way to make it real and to document what is happening. Cities often don't even have access to their own data. They'll have legal services, representatives, lawyers, public defenders who come to them and say, “Hey, lots of evictions are going on.” And they'll say, “Oh really?” Because nobody's had the data. (For many years, it was all paper.) Nobody could track evictions at all. This research is showing people that it is actually happening. We are starting to show the non-believers.
Progressive cities— New York and San Francisco—have always known that there's tons of evictions out there. They work to have tenant rights to counsel and other policies to prevent it. But, for instance, places like Texas have been incredulous that it's actually a problem. So maybe if they start to see the eviction numbers go up in Houston and Dallas, some of these other cities will get the idea that this is actually a real crisis.
IGS: I understand your project with the state of California addresses issues related to populations hardest hit by the pandemic.
KC: Yes, the state has a number of different strategies that it is trying to put in place to respond to the pandemic. One of their dreams has been to implement Project Roomkey, which converts hotels to housing for the homeless. There's actually a big data problem behind that proposition. There's no central listing of hotels in the state of California. You can go to individual cities and ask them for their list of hotels, but then how do you figure out who will be a good candidate to house the homeless? So one of the things we're doing for them, among many other things, is we're using big data, like Yelp data, to identify potentially good candidates for participating in these programs. Through Yelp and other kinds of online data sources, we can figure out how high the occupancy rates are for hotels, what their cost structure is and so forth, to identify some candidates. Then the state can reach out and make an offer for them to participate in the program.
IGS: I've heard of the frontline health worker initiatives, where places like New York City have established online portals that update the availability of hotel rooms for frontline workers in real time. Would it be possible to piggyback off of their model and use the information they've gathered and apply it to initiatives like Project Roomkey?
KC: That's right. There is a close connection. I think San Francisco did that. They reserved a bunch of rooms for essential workers and ended up using them for the homeless.
IGS: That’s hopeful.
KC: And then with our Housing Precarity Risk model, we will link this data to California specifically, and identify areas that are most in need.
IGS: I’d like to ask you a personal question about doing research during a pandemic. How are you managing to organize your team and communicate with them effectively? How has the pandemic impacted the logistics associated with doing your research so far?
KC: Oh, well actually, we're all much more productive than we usually are. It’s pretty amazing. We've never really been able to concentrate on campus. Our space at Anna Head had a black mold problem. Our team left a couple of years ago, so we haven't been working in a shared physical space in a long time. We've been using Zoom for many years. And then our lead researcher had COVID and has worked through the whole thing. They were very sick with a fever for two weeks over in Oakland.
IGS: Oh no. I am so sorry.
KC: It’s pretty amazing that that’s the only person I know who’s actually had it. You know, in some ways, we are closer now. We probably meet more often because we're completely on Zoom and not doing in-person meetings. There are no constraints. We use Slack. I think lots of teams like ours have been very fortunate in that we are all set up for this kind of communication.
IGS: We are the lucky ones – to work at the university in a situation that doesn’t require close contact, that isn't interfacing with a lot of the public.
KC: It's really brought home how privileged we are.
IGS: Yes, absolutely. Thank you so much, Professor Chapple, for your work on these projects which address aspects of the pandemic as critical as housing, and for showing us who is the most impacted. We look forward to seeing where the projects go and what you find.