In this episode, Stephen Goldsmith interviews Carin Clary, director of Homelessness and Housing at Harvard's Government Performance Lab, about the challenges faced by cities in addressing homelessness, the role of data in guiding effective interventions, and innovative approaches to reshaping homeless response systems.
In this episode, Stephen Goldsmith interviews Carin Clary, the director of Homelessness and Housing at Harvard's Government Performance Lab. They discuss the challenges faced by cities in addressing homelessness, the crucial role of data in guiding effective interventions, and innovative approaches to reshaping homeless response systems for lasting impact. Clary offers valuable insights into data-driven strategies that not only tackle immediate challenges but also provide a roadmap for cities seeking sustainable solutions to this cross-sector problem.
Music credit: Summer-Man by Ketsa
About Data-Smart City Solutions
Data-Smart City Solutions, housed at the Bloomberg Center for Cities at Harvard University, is working to catalyze the adoption of data projects on the local government level by serving as a central resource for cities interested in this emerging field. We highlight best practices, top innovators, and promising case studies while also connecting leading industry, academic, and government officials. Our research focus is the intersection of government and data, ranging from open data and predictive analytics to civic engagement technology. We seek to promote the combination of integrated, cross-agency data with community data to better discover and preemptively address civic problems. To learn more visit us online and follow us on Twitter.
Betsy Gardner:
This is Betsy Gardner, editor of Data-Smart City Solutions at the Bloomberg Center for Cities at Harvard University. And you're listening to Data-Smart City Pod, where we bring on the top innovators and experts to discuss the future of cities and how to become data smart.
Stephen Goldsmith:
Welcome I’m Stephen Goldsmith. I'm a professor of the Practice of Urban Policy at the Harvard Kindy School, and we have another interesting guest on our podcast, Carin Clary, who's the director of Homelessness and Housing at Harvard's Government Performance Lab, and she leads a team that supports state and local governments. Carin, welcome. Thank you for being with us. Before you get started, tell us a little bit about your journey to Harvard. What were you doing before?
Carin Clary:
Thank you so much for having me. I'm really thrilled to be here, and I am currently the director for Homelessness and Housing at the Government Performance Lab. Right before the start of 2022, my return to the office after maternity leave was going to coincide with a new mayoral administration. This being my third that once again, homelessness was really high on the docket of priorities, but this time with heightened concerns of public safety and budget deficits in the intermix. I had so much angst as to what this would mean for my team in terms of pressures to reduce homelessness and finding ways to do more with less, and I spent a lot of time thinking about what meaningful supports to staff would look like. I had spent nearly 20 years in New York and a lot of vantage points in the public sector, from mayor's office to discrete agencies to nonprofits, and New York in many ways is really the high watermark in terms of just local infusion of dollars into the public programs and also the type of talent that it could attract to its workforce.
Where I landed in my thinking was that homeless response is an emergency system that is always an emergency, whether that's human emergencies of life and safety, political firestorms, natural disasters, you name it. And when that emergency is not short lived and it becomes the sustained normal, it leaves so little room for thinking about innovation, trying to sustain a really constant emergency. Really for me, always produce three results. Number one, I would use my best and brightest staff to handle those situations. Number two, I would just inevitably fall back to doing the same systems, but just try harder to squeeze more out of them. Or three, sometimes I'd be able to convince a boss or a procurement team to allow me to bring in external support. The times that I did bring in external support and I worked with really great consulting or TA firms, I really found that if you don't have the capacity to absorb that TA or the capacity to actually implement the smart suggestions, there's really very little return on that investment.
For me, what often would happen was that we would bring someone in, I would have to dedicate a lot of my time or my team's time to really educate the folks on the problem and share ideas of what I already knew or thought about what was needed, and then that would produce a very beautiful glossy report that would sit on my desk and I'd say, someday I will have the time and space to put those smart ideas into action. But that really never happened. So when I came across the model of the Government Performance Lab, it immediately resonated as something of value to me as a former public servant, so much so that I was ready to step away from local government and pursuit of bringing people support that I thought really could make a meaningful difference. I would say the one thing that is sort of different about our model is that we don't just provide capacity that can get pulled into the emergency du jour, but we try to be really specific in working shoulder to shoulder with staff to lift up what the existing staff ideas are about what's needed, and then we really try to provide support that can be the doers of actually implementing those change, and we really try to be focused on testing things that not only have value for that community or that government, but really are lessons that we think will have national resonance in terms of their application.
Stephen Goldsmith:
Homelessness is a big issue. We've written about it a lot, and we run these programs for chiefs of staff and chief data officers, and it's a frequent subject. We're not going to solve it in 15 minutes because it deals with so many things, zoning codes, housing costs, housing stock, public health, healthcare, criminal justice, addiction treatment, et cetera. But why don't you help us focus about what specifically you're doing with localities that might make a difference for homelessness, and then I'm going to spend most of our time today talking about the use of data as part of the solution set.
Carin Clary:
Yeah, I think that's exactly right. I think there's a bigger underlying question in that of who's really responsible for homelessness. So much of the narrative is so singular, the perception that it's an individual personal set of circumstances or it falls on those responding in the literal moment of homelessness. But in reality, we know a lot more folks at the table when we think about the problem of and the solutions to homelessness. The current homeless response framework predominantly focuses on just the literal point of homelessness and puts really an outsized charge on the singular actors or the agencies that are managing those emergency systems. When in reality, homelessness stems from failures in major upstream and social issues such as poverty, historically, racist housing policies, critical shortages in housing availability and stock, and huge things like access to healthcare and the larger behavioral health system agencies are extremely limited in their positionality to access the real things that would make a difference, like the creation of more deeply affordable housing, flexibility in how they do contracting access to even eviction data, and yet they're still charged with making meaningful quantifiable progress on this huge issue.
Moreover, outside of big places that have a high infusion of local dollars dedicated to homelessness like New York City like Los Angeles, San Fran or Seattle, there usually isn't even a designated agency that is doing homeless response. You have a small team that is within one agency or the responsibility is diffused across a number of community stakeholders that are known as continuums of care, the Government Performance Lab or the GPL as we use a shorthand. We recently did a call for applications to use our technical assistance through our homelessness prevention and rehousing accelerator, and for just two TA slots, we received over 100 applications from all parts of the country, both big and small. There were numerous jurisdictions that wrote in things like 'I am the only person in my entire county that is designated to work on homelessness. Please help.'
Stephen Goldsmith:
So much stuff here. Hold on a second. I've got a bunch of questions I want to get to data, but let's go back to your key thesis here that often no one is in charge. In an ideal situation where you're doing prevention as well as intervention, where should that jurisdiction be located? To whom should that person report and how do they organize the fragmented system of city services to produce a better solution?
Carin Clary:
Yeah, no, it's a great question, and I think it does actually link to the question of data because I think data can be one of the most important tools in terms of thinking about how you build that transparency and how you bring that accountability around. I think we have to really think about how we can either have either a strong principal, maybe at a mayoral office that is thinking about how they are taking that mantle to really bring other agencies to the table to think about how they're contributing to the problem. Either things like how institutions are directly discharging people into homelessness from criminal justice, from child welfare, from sort of healthcare to thinking about the ultimate solutions, which is much more around the actual creation of deeply affordable housing and where people are in terms of dedicating actual units that will be sort of purpose built.
I think that can look anything from taking a very hard line, like a CompStat of really bringing people together and looking at real-time data to see where our entrance to homeless coming from and from what institutions and how do we translate that data to actual actionable insights and change in discharge policies or how we think about people coming out of institutions and not having to fail further into homelessness before they get assistance. Or it could be things that are more like how we think about prevention and how we can actually track the different points that people are experiencing, housing precarity, anything from calling in for lighter touch services all the way to being able to get right to counsel and eviction support. I think it could be anything on that continuum of taking a very hard line, too much more maybe softer approaches of really thinking about how do we elevate the question of housing stability as being a shared focus, not because you're being compelled to, but really thinking about how does housing stability influence some of their other priorities that maybe are at the top of their mind with questions like community safety around questions around housing stability for families and for children that have experienced involvement in other often punitive systems.
Stephen Goldsmith:
Okay, occasionally you have to take a breath, so I can ask a question. A few years ago, I saw with Mayor Garcetti in LA maps, I think they've been done by Esri, but maps that showed all of the above the things you mentioned, it showed where homelessness and lack of housing was occurring, where evictions were occurring, where treatment resources were occurring, all of the layers of the issue that could drive decision making. So what role do you see for maps and B, are there data platforms that allow multiple agencies to work together with a kind of common visibility into the problem?
Carin Clary:
I think part of the paradox that I think we've created nationally is that we really want to focus our extremely precious resource of deeply affordable housing to only go to those that are most vulnerable, those that are most in a need, and in doing so, we protect those resources by making the burden of proof extremely high for eligibility, and then we in sort of government get in our own way. I think one of the biggest challenges to the system is that everything is still run so manually and so sequentially that the process to connect to housing, especially subsidized housing, has way too many steps and it's unnecessarily burdensome for people that are experiencing homelessness as well as government staff that are trying to support them. I think the very manual process and the reliance makes it extremely inefficient and how we refer people to housing, but also makes it incredibly difficult to track critical data points, both thinking on an individual client level and also system level.
So just to give you an example, if you wanted to get a sense of where a client is in the process before an official outcome has been hit, you likely have to dig through just trails and trails of emails, and then you got to put that together in the aggregate. It would be the same process. I think one of the key issues in homelessness is that in most places, there isn't even real-time data about who is experiencing homelessness. Most places of the country, they are getting their primary funding for homeless response through the federal government, which requires them to do a once a year point in time count, which is incredibly important and has become reliable because it's been used for so many years, but it's really just a snapshot and in fact, we actually need to have much more actionable data to really be able to effectuate change
Stephen Goldsmith:
Effect. Yes. Let me ask you a question. We've had some presentations, some about housing supply and its effect on homelessness, some about the Houston model that's gotten so much attention. We wrote a book, Kate Coleman and I a couple years ago on cross-sector collaboration. One of the chapters is on homelessness. We interviewed some folks in Santa Clara County who were looking at those homeless folks who found themselves most frequently picked up on the street. Most frequently they're taken to jail or mental health intake facility, and then focusing their resources on supported housing for those who had the most frequent visitations on the theory that there's high return on investment. Do those models make sense? Have you seen those models other places?
Carin Clary:
Yeah. Actually, one of the first housing projects that the GPL ever ran did just that looking at high utilizers of multi-systems such as ERs and jails and shelter to demonstrate the value of, in this case, permanent supportive housing for targeted groups of unhoused people. That program is actually in its third year of an RTC evaluation, and it's consistently been shown to have positive impact on housing stability as well as decreases on the reliance of emergency systems today. One of the projects that we're actually doing in Colorado is on a statewide level that really works with the state's Department of Corrections that oversees the prisons really to work to identify those that are at risk of homelessness upon being discharged and using the time pre-release more effectively and more firmly rooting housing related services. I think that homeless response is often too narrowly focused on the moment of actual homelessness, and then we are often waiting until a bad outcome happens, which leaves opportunity on the table in the form of time, and specifically even just interactions that they're already having before a person is released. I think when we interviewed the state of Colorado, when they were a contender to receive TA from us, what they said to us was that we are incredibly data rich, but we're very capacity poor. So with the extra support that we bring from the government performance lab, we're really focused on helping them make better use of the data, realigning how they do discharge services, and ultimately being more efficient with the time so that people are not being discharged from prison and then having to fail further into homelessness before they can get help.
Stephen Goldsmith:
Did you intentionally not, or did I miss the name of the city where you said you had done the work with frequent utilizers? It was...
Carin Clary:
Denver. Denver, Colorado. Yes. One of the first ones we did was in Denver, Colorado. That was a number of years ago, and then today we are currently running a project that is statewide out of the governor's office in Colorado.
Stephen Goldsmith:
And you have a statewide project in New Jersey, as I understand too. What's that involve?
Carin Clary:
Yes, yes. We're really excited about this. When we think about New Jersey, one of the key things is that I think within homeless response, most people that are working there have been trained in a very direct service capacity. So they're really great at coaching individual people, but they've often never been given training at looking at cases in the aggregate. I think we're asking workers to both be experts in working with individuals that may have a very wide and far range of needs, including trauma, dv, behavioral health, and we're asking them to be systems thinkers, but we really only invest in half of that very tall order. We've seen that a lot of jurisdictions are doing a good job of articulating priorities like wanting to increase exits from shelter or reducing the number of people that return to homelessness, but really where it breaks down is how they achieve that goal.
To be able to address any of those goals, you would really need to be able to break down all of the housing and of the placements and then use that data to troubleshoot where the system isn't working and where it's having success. But in order to actually do that operationally, government workers would have to be able to use Excel. They would need to understand how to spot meaningful trends in the data, and then on top of it, they need to come up with an actual plan to address it. Most often what we see in jurisdictions is that there's a scenario that looks something like a single set of data is shown. Let's say it's showing that exits from shelter have gone down and then the deciding body sort of switches into a lot of anecdotal accounts about what's happening, or there is a panic that results in the announcement of a task force to address, or there is some arbitrary sort of commitment that's made about decreasing homelessness by X amount.
Also, in that scenario, we see that a lot of people that have lived experience have been brought in to act as an advisory capacity and to translate what their qualitative lived experience is, but there hasn't been a lot of thought as how you do that to support an authentic engagement. So the state of New Jersey really did two important things. One, they invested in creating designated data analysts, staff positions that didn't previously exist to support half of the state, and two, they invested in training for the leaders in the continuums of care that don't necessarily have a designated data function, but to really support them and how they're using to really lead to organizational change. This includes people that are executives of nonprofits and people with lived experience. So it's a really wide array of people that are in the room for us. There's admittedly a very wide array of stakeholders when you think about the data analysts, executives, and people with lived experience, and then when we think about how we're trying to upskill them, and there's huge variance just in terms of their data literacy, their comfort and training settings, and frankly even just in their proximity to power, but that is really how homeless response functions in most places of the country.
In most places, they're just getting that federal investment, which then requires them to sort of loosely organize locally, and it's a combination of some government agencies, faith-based organizations, community organizations, and people with lived experience that are then sort of charged with coming together to make decisions around how they're using funding and sort of what their local priorities are. We were admittedly really nervous to think about how we were going to cover this broad range of an audience, but I think now we're really excited to think about this as a microcosm of what homeless response looks like nationally and really trying to understand what the needs are and what it would take to move the needle.
Stephen Goldsmith:
So what's the relationship between those state data analysts and the city of Newark? I don’t understand quite how they support local interventions.
Carin Clary:
The way that it works is that in most of the parts of the state, there are sort of different regions that are organized as continuums of care, which will include big cities like Trenton or sort of a Newark or sort of other parts of the state. We're training half of the COCs in the state, so half of those regional groups that are a mixture of government people, but largely service providers that are coming together. And so this will be someone that sits within one of those agencies or one of those organizations, and they are providing support not just to that organization, but to the broader COC. So they're looking at data about people experiencing homelessness, not just for one organization or one agency, but for all people that are being served by that region.
Stephen Goldsmith:
I know we're running out of time. What two or three specific recommendations would you make to cities about how they would set up the data capacity for better decision making, intervention prevention with respect to homelessness?
Carin Clary:
I think people really need data in four ways to make meaningful progress on this. I think on an operational level, they need to be able to get access to data to both prevent homelessness, understand where people are coming into the system from what is successful in terms of being able to prevent homelessness. They too, they need data to be able to understand people in real time, not just once a year, that are experiencing homelessness since it changes so much over the course of the year and really how they are being connected to different services. Three, they need reliable data about the actual housing stock. Most places, if you needed to know what the vacancies were in a program would have to call around to each of those housing programs and sort of tally that. So they need reliable, actionable data on vacancies and four, they need data to be able to really evaluate what's working.
Most places, they rely on the federal system, which is really set up around compliance, and it is not set up to be something that is easily manipulable that you can extract data that you can use for operational insights. So investments in either how to make that work better or systems that can sit on top or alongside it are huge. But I think largely it's around staff capacity. I think there is a huge possible crisis around human capital and loss of retention. When we were doing our call for applications and sort of thinking about our programs, we spoke to people all across the country. Of course, the first thing they're saying was still largely about the demand outpacing the supply, but the second thing that they said was really like, we do not have staff. We cannot find staff. We cannot retain staff. We are buckling from within. And so I think there is this critical moment to think about how we are investing in the workforce to do this incredibly complex hard work and thinking about how we are sort of seeding the future workforce to be sort of leaders when we know that homelessness is not something that is going to go away in any sort of immediate future.
Stephen Goldsmith:
I know when you're finished, of course New Jersey will be the best of the pack, but right now, what one or two cities would you think others should look at concerning the best use of data as part of the response to homelessness?
Carin Clary:
Well, I'm a little bit biased because I'm going to say programs that we're currently working with, but thinking about prevention. We're really excited about the work that we're doing in Detroit, which is setting up a single access point for anyone that needs help with housing to call in. That is providing us with just a wealth of information to understand who calls and when and why, and then being able to sort of track how successful each of those interactions are in terms of the resources that they get referred to. So we can learn a ton about how people are moving through housing precarity and not just a literal question of where were you last night before you entered the front door of the shelter system. So I think that is incredibly compelling, especially when you think about in most places, we only have information after the bad outcome has happened and a person is homeless.
To be able to get meaningful data that's further upstream and really understand the complexity of people's situations is huge. On the housing side, I think the thing that I love to talk about, and the drum I always beat, is really around the process and how manual these processes are. Permanent supportive housing, because it can only go to people that are homeless and has built in behavioral healthcare supports that are there. It is our most precious resource, but the way by which people get connected and the rules around proving eligibility are way too high. It often can take 3, 6, 9 months before a person is actually moving into a unit. I think the work that we're doing in Chicago to really think about how do we collect more information, more data upfront about people's eligibility, about their preferences, really will allow to have a higher success rate, which means you don't have units sitting empty because we can operate just more efficiently, effectively, and equitable with one of our most precious resources for those that need it, the most people that are chronically homeless with behavioral health.
Stephen Goldsmith:
Well, as you can tell, Carin has lots to say and lots to help cities with concerning homelessness. We'll continue to direct folks to you and your center. This is Steve Goldsmith. With me today is Carin Clary, the director of Homelessness and Housing at Harvard's Government Performance Lab. Thank you so much for your insights and your time.
Carin Clary:
Oh, thank you so much for having me. It's been a pleasure.
Betsy Gardner:
If you liked this podcast, please visit us @datassmartcities.org. Find us on iTunes, Spotify, or wherever you get your podcasts. This podcast was hosted by Steven Goldsmith and produced by me, Betsy Gardner. Thanks for listening.