The Cost of Homelessness in America | Dr. Dennis Culhane

Tristan Markle (00:00)
This is Policy Crimes, a podcast about the price of bad policy and the value of doing the right thing. I'm your host, Tristan Markle. Across the major policy challenges of our time, we see the same pattern: governments are spending more to fail than it would cost to succeed. It often costs more to keep people homeless than to house them, to keep child care expensive than to create affordable options, or to subsidize fossil fuels than to invest in green energy.

But Policy Crimes also looks at the alternatives, ideas that can save both money and lives. In this first season, or case file, we'll be investigating the cost of homelessness. Governments across North America are spending billions managing homelessness, but the crisis keeps growing. The research is clear though, providing permanent housing isn’t just more effective, it's often less expensive. So why aren't we doing it?

This season, we dig into that question. And over these episodes, we explore both the cost of doing nothing and the value of doing the right thing.

To launch this season, we begin with Dr. Dennis Culhane, professor of social policy at the University of Pennsylvania. Starting in the 1990s, Dr. Culhane's research revealed that governments were already spending tens of thousands of dollars per person per year to keep people homeless through shelters, hospitals, jails. His work helps us understand who becomes and stays homeless and why. And he's laid the foundations for cost of homelessness studies while shaping major housing first policies across the US. In this episode, we trace how Dr. Culhane has challenged old judgmental narratives about homelessness and reframed it as a solvable issue where solutions could be more cost effective than the status quo.

Tristan Markle (02:18)
Welcome Dennis Culhane to our first episode on the cost of homelessness. We know that you've been really involved in homelessness research for a long time, including some really important work that led to some of the early cost of homelessness studies. Going back to that early research, could you take us back to the early 90s and what were you trying to understand when you started looking at that shelter data in New York City and Philly?

Dennis P Culhane (02:46)
Really, there were two major gaps in the literature at that time, I would say. One was there was a numbers debate, as it was called, about how many people actually experience homelessness. And all the measures we had were based on cross-sectional surveys on a given day. We sometimes use the term a PIT count, a point in time count. Researchers didn't really think that that was problematic because there was an assumption that the population didn't change very much from day to day, that it was the same people who were there day after day after day. So a one day measure was good because you didn't double count people, which was the main concern, and that it probably fairly represented the size of the population.

But there were countervailing views as well that, indeed, maybe there's a lot more dynamic turnover in this population than people expect. And the longitudinal data that we got in New York and Philadelphia provided the first opportunity to look precisely at that question. So if, for example, Philadelphia has like 5,000 people in a shelter on a given day, how many are going through that in the course of year? It turns out to be about 15,000 or 16,000 people. So we saw turnover rates of three to four, showing a much larger population. So that number was the first thing.

But the second thing was there was a conceptual model out there that there were three types of homelessness. There was transitional homelessness, short-term homelessness. There was repeated episodic homelessness. And there was chronic homelessness. And people had talked about this, but there really had never been any attempt to measure it and to see how big were these relative populations. What the longitudinal data revealed was that, in fact, the vast majority of people, 80%, are in this short-term category, who are using the system for relatively brief periods of time, like less than a month on average at that point.

And then there's a small group of people, 10% who are chronically homeless, but importantly they use half of the total bed days in the system, so that even though they're 10% of the people coming into homelessness in the course of a year, if you look at the shelter beds on a given day, half the beds are occupied by that 10%. So that's why it's possible for people to think the population doesn't change very much: we see these same people here day after day after day. Because it's true, about half the population is there on a long-term basis. What you don't see is all of the people who sort of are anonymously, invisibly passing through the system. They don't look homeless. They're not recognizable as quote unquote homeless.

And so it revealed that there were these very distinct populations. And to your point about utilization, the main takeaway from the shelter manager policymaker situation was that most of our resources are going to serve a relatively small group of people.

And we're doing it in very threadbare kinds of accommodations, you know, a cot that is maybe 15 to 20 inches from the other cot, and in these large rooms, and there’s very little privacy, and it costs a lot of money. Apartments are definitely cheaper than beds. So there was kind of a realization, wow, if we took these folks who are long-term shelter clients and got them into housing, it would probably be cheaper in terms of the cost of the housing and we would be able to cut the number of beds that we're paying for by half. So that was the beginning. I think that was the first insight that kind of turned things upside down a little bit.

Tristan Markle (06:38)
There's kind of two insights there. One, that 10% of the people, we were spending 50% of their resources on them. And then the other side is that because the system was putting more resources to them, maybe we were paying more attention to them and not noticing that instead of 5,000 people homeless on a night, there was 15,000 to 20,000 people homeless in a year, and we weren't appreciating that. When you did that study, were you surprised to find that for 10% of the clients, we were spending 50% of our resources keeping them in shelters? Was that a surprise to you?

Dennis P Culhane (07:14)
I would say it was not a surprise to me, because I had been doing research in mental health with my colleague, Trevor Hadley, who was a psychologist and had run state mental health systems. And he had popularized an analysis called heavy users of inpatient psychiatric services. And what he had found was that 5% of the serious mentally ill population accounted for more than half of all of the hospital costs for people with severe mental illness. So this phenomenon of a relatively small group of people dominating the use of resources is actually something that you see in other systems. You see it in local jails. You see it in mental health systems, for example. I would suspect it's true in some of the substance abuse treatment systems as well.

And you've also probably heard this 80-20 rule that people talk about in business. 80% of your customers account for 20% of your resources, or the flip side, 80 % of your resources are really coming from 20% of your customers. So there's been this sort of idea of the repeat customer being the dominant user of resources or spender in your system. And it's common in other systems as well.

Tristan Markle (08:37)
What was the response from others to that initial study, other researchers or policymakers or officials?

Dennis P Culhane (08:45)
Well, I was initially focused on the fact that it demonstrated that homelessness was mostly an economic problem because so many people were experiencing this condition, but on a brief basis, and they were getting out on their own. You know, most of them returning to employment, getting back into housing, what have you. So to me, it was a counter argument to the pathologizing of homelessness that had been going on.

But when I started to present these results to people like the city of New York, the managers who ran the shelter system in New York, or here in Philadelphia, they immediately focused on the fact that most of the money was being spent on a small group of people and that could be done much better. So that the whole idea of having a chronic homeless focus policy really emerged from those initial conversations.Places like the city of New York and the mayor's office who I was working with at the time, immediately focused on that. So even though my original article, if you go back and read we talk about that, but we also lean heavy on the fact that homelessness looks like it's mostly an economic short-term problem.

Tristan Markle (09:58)
Before we move on to the cost of homelessness, I have heard you say that the most common length of time that people were in a shelter was one day and the second most common was two days. So could you speak to that? Because that's an important insight.

Dennis P Culhane (10:11)
Sure, yeah, I think that most people don't appreciate that people who experience homelessness are just like everybody else. They don't even use the word homeless to describe themselves. They just know that they're without a place to stay and they somehow find their way, where can I go? And they end up in these places that are called homeless shelters, but most people don't use that language and they don't identify with it in that way.

But like any of us, if you ever had to stay in a shelter, the very first thing on your mind is how do you make sure you don't have to go back there the next night? Because the conditions are so difficult, it's almost impossible to sleep, it's loud, you're in this big facility, you have no privacy, you're concerned for your safety, you're concerned that your things are gonna be stolen, and there's just no redeeming aspect to the shelter other than you're not sleeping outside, possibly.

So people really are highly self-motivated to get out of homelessness. That's what that shows. And the curves consistently show the same thing, that the most frequent time is one day followed by two days, and it goes pretty steeply down, so that about half the people, you get to the median, I think I just saw the most recent data, you get to the median length of time homeless duration is now somewhere around 50 days.

Tristan Markle (11:33)
Yeah, and I think we see that in Canada and the United States, a similar number.

Dennis P Culhane (11:37)
Yes, these typologies that we had set out have been replicated throughout Europe and in Canada and the United States.

Tristan Markle (11:46)
Moving on to your first big cost homelessness study, you published in 2001 or 2002, you were setting out to measure the relative cost of people who are homeless versus people who had been homeless but were subsequently housed, and what were the costs on the government services that they used before and after. And you looked at 10,000 people with severe mental illness, so it was a subgroup, is that right?

Dennis P Culhane (12:17)
That's correct, yes. Everyone in the study had a severe mental illness and was homeless at the time of enrollment. That's right.

Tristan Markle (12:26)
What did you find?

Dennis P Culhane (12:29)
First we just matched up those records of the persons with their name, their birth date, social security numbers, and we matched them up with different data sets we could get access to. The primary data, of course, where we found them was from the shelter records. And then we matched it up with state psychiatric hospital data, community general hospitals paid by Medicaid. New York City has a public hospitals corporation, the Health and Hospitals Corporation. We had outpatient Medicaid. We had VA hospitals. And then we had local jail and state prison data. So we had these eight data sets, and we were able to look at what was the use of services across this 10,000, before anybody got into housing, looking at this 10,000 group of people, while they're homeless, what kind of use of services did they have? And we found that basically almost every other day, almost one out of every two days, they were in one of those institutions.

Primarily the biggest one, of course, was being in shelter. On average, the person was in shelter for four months out of the two year period that we looked. But they had two months in a psychiatric hospital, a couple of weeks in the public hospital system, a month in the community general hospitals. So, a lot of heavy healthcare use, about 85% of the dollars were associated with healthcare use. And they were in one of these healthcare facilities almost one out of every four days. So essentially, although people might think, homelessness, these people don't cost anything, they're just on the street, they're not living anywhere, they're not using services. The fact is they do end up hospitalized, very high rates, they end up in shelter.

And the total average cost per year in 2024 dollars is about $70,000 per person per year. So that's a tremendous public expense. At the end of which, none of these programs actually end people's homelessness. They're served in these facilities for the purpose of those facilities, then they're discharged. So you're spending all this money and people are still homeless.

But as part of the study, we did have this big intervention, the New York-New York housing intervention. It was a big experiment to get chronically homeless people with mental illness into supported housing. About half of that 10,000 population got into this housing. And when we looked at the use of services, they cut their service use by about $28,000 in 2024 dollars, a substantial reduction, more than a third approaching close to half actually, of the resources we were spending because they were no longer using those services. And then when we compared it to the cost of the housing, it was $30,000, it was effectively break even. 95% of the cost of the housing and the support services, the case management, et cetera, was accounted for by reductions in these other uses of acute services.

That was a pretty big important insight – that keeping people homeless was costing taxpayers an enormous amount of money to no good benefit, and that shifting to a housing focus didn't really increase overall costs to the taxpayer. And that you got this much better outcome. People are housed, they're stably housed, they're not living on the street. I mean, what's not to like about that story?

Then that proceeded to then lead to the Chronic Homeless Initiative in the Bush administration.

Tristan Markle (16:22)
Just to summarize for everyone, in that study of 5,000 people who remained homeless, on average, governments were spending $70,000 a year on each person, And then when they were housed, you were spending about $70,000 a year, but housing them. You had two options, house them or not. And it cost the same amount of money in your study. Or more, because you didn't count ambulances and some other costs.

Dennis P Culhane (16:50)
Well, that's true. That's true. The study has been replicated many times, but one of the things we learned was that there was this fairly costly item that other researchers found, which is the use of EMT services. And that if you included those, we would have got better than breakeven. Our study did not include EMT charges, but we know that most people who are homeless who get to a hospital are usually taken by EMT, and that can run six to eight hundred dollars a run.

Tristan Markle (17:23)
What were the responses to that study? Probably a lot of activists said, well, that's what we've been saying for a long time, but maybe government wasn't believing them.

Dennis P Culhane (17:34)
Right, people expect advocates to make claims like this. And in fact, there were signs in the New York City subway system, on the subway cars, every year you would see these signs about, you know, a jail cell costs X dollars, like $70,000, a shelter bed costs $50,000, housing costs $20,000. There were things like that in the subway. But again, these are the kind of arguments you expect advocates to make, and there's a difference between research statistics and advocacy statistics. So this was really the first time that we would say independent analysis by people who were not in the industry, so to speak, but just separately looking at it. So I think it carried more weight in terms of validity in that regard.

But the reaction was pretty big. There was a pretty decent industry forming in New York City around supported housing. So they were very excited to be able to promote these results to their funders in government. There were government agency leaders who also were very attracted to the positive outcome without huge increase in spending.

But the challenge was that the money that goes into shelter, the money that goes into hospitals, you can't just take the money from there and put it into these housing programs. So it created a whole conversation that has gone on for years about how do you front load the investment? Where to get the dollars so that you can achieve these cost offsets? And then the people who save money, like principally in our country here Medicaid is the program where most of the savings accrue. But yet Medicaid really can't pay for housing, can't be used to pay rent. So there has to be a higher level of government at the executive level where the executive branch can say, we can take the money from here, invest it there, and it will save us money on the Medicaid side. You need somebody, a government official who has that perspective and who can then make that neutrality argument because trying to get the healthcare agency to pay for housing has been very, very difficult.

Tristan Markle (20:03)
And so this cost offset issue is something that we were really hoping to dive into through this whole podcast series. Could you explain for our listeners just what the term cost offset means in everyday language? Because I think part of it is a translation issue.

Dennis P Culhane (20:24)
Right. Economists, they study cost effectiveness, cost benefit analysis. Those concepts involve other value judgments regarding the benefit that might derive or the effect that might derive from certain expenditures. A cost offset analysis is really just more of a cost accounting framework to say, okay, we spent this much over here, what is the secondary impact of that on other services? How much does that reduce, or is the cost of that offset, by other programs? So it's a more technical sort of audit-oriented statement that you can make regarding what was spent and what went unspent as a result of that. And that's basically the cost offset.

We're not making the argument about cost benefit analysis. That's a deeper thing. And really, there are more judgments that go into the decision of what to include and how do you monetize the value of a benefit of someone having a place to live, or the quality of life years associated with being stably housed as opposed to living outside? There's a bunch of these other measures, and I'm not an economist so we just went with the straight calculations we could do as quickly as we had in that study.

Tristan Markle (21:37)
So let's break down just that offset analysis a little bit and some of the challenges with it that you already brought up. So one problem might be what we call the fixed cost problem. You kind of mentioned that, that just because there's cost savings – people who are homeless are using fewer hospital beds – now that doesn't mean you're going to cancel the hospital beds or stop funding them.

Dennis P Culhane (22:02)
Yes.

Tristan Markle (22:13)
It does mean that other people can use them and it's more efficient. But nevertheless, there's a problem getting governments to acknowledge or account for that savings. So that's the fixed cost problem. Another one, second one is the intra or interagency silos. So how does one government department agency like healthcare account for cost savings that are produced by another agency's housing intervention?

Dennis P Culhane (22:24)
That's Correct. Yes.

Tristan Markle (22:41)
And then there's different jurisdictions, in the US you have states and federal governments. We have provinces. I just want to quote to you one of your own quotes about this and see how you reflect on it 20 years later. You wrote, this is the fixed cost problem, “In systems where services are funded by direct support or subsidy of facilities and operational activities like jails and not through cost-based reimbursement systems as in healthcare, reduced utilization by some people will not reduce the overall facility operating costs as those costs are paid irrespective of who uses the system or for how long.” But as I just mentioned, it’s not too hard to think about the future demand that you are able to address, so why can't we account for the future demand?

Dennis P Culhane (23:32)
I think that that's a very good point because in that quote, particularly, we were looking at the time horizon of this cost accounting across the one or two year period. But as we've seen here in the United States with the number of reforms that have happened in the criminal justice system in terms of decriminalizing marijuana, for example, changing the sentencing for powder versus crack cocaine, but also significantly bail reform, we've seen very significant declines in the use of jails in the United States over the last five or so years, because many systems have gotten much smarter about who needs to be detained and who does not. So it does speak to the fact that on a longer time horizon, you can have a more reform-oriented mindset where you can say, we can actually plan on fewer homeless people being in our jails if we're gonna instead have them in housing. And we may not see the difference next year, but we can see the difference a few years down the line.

Tristan Markle (24:43)
Exactly. In your 2002 study, you measured jails and hospitals and other institutions. And it's not uncommon for health economists to measure over a five-year time frame. I mean, that's the most common, isn't it?

Dennis P Culhane (24:48)
Yes, right. And healthcare, as stated in that quote, healthcare is a cost-based reimbursed system. So if someone's not in a hospital, then our Medicaid is not paying for that. So there's an actual reduced cost there. It doesn't mean they don't have other demands elsewhere. And we do have a shortage of inpatient behavioral health beds. So empty beds tend to get filled. But that's really just because the system isn't really operating to meet the demand as it exists out there. So it’s a good point.

Tristan Markle (25:34)
Are there any examples of any level of government or any of these agencies accounting for cost offsets in their budgets or policy and legislation?

Dennis P Culhane (25:43)
Not like I would like. I don’t remember the exact example, but there have been jails that have closed because they were no longer needed. But very specifically within Medicaid, which is our primary health insurance for low income and disabled elderly people, Medicaid requires that if you're going to increase expenditure in some category of services, you have to demonstrate the neutrality of that, or near neutrality of that, within 12 months of that service. So whereas, in this case of housing, it might take two years before a person's health has been stabilized such that now that they're in housing, you see the offsets much more equalizing. So it might take two, might take three years.

Tristan Markle (26:43)
Or even longer with the prison system where it takes years for the trials to go through.

Dennis P Culhane (26:50)
Yes. So our health care system is not attending to that now and you could argue that the private health insurance companies that now administer this Medicaid money, they could take a longer view, and they should be taking a longer view. But the official waivers that you get from Medicaid to enable you to pay for housing-like activities, they want to see a cost offset in 12 months, which is unfortunately, I think, too short-sighted.

Tristan Markle (27:21)
I don't know enough about the American system on how to get them to extend that time frame.

Dennis P Culhane (27:26)
I know, it's very frustrating. But you know, there could be certainly more imaginative people thinking about this, especially the uniqueness of this population. It's widely understood that we have a system in healthcare, for example, where 5% of the patients served by Medicaid account for 65% of the expenditures. That's that heavy user phenomenon. And one of those five, one out of every five of those people, is someone who's experiencing homelessness. So homelessness is well known to be one of the principal drivers of very high expenditures in healthcare. Others are of course people in nursing homes, people with very severe illnesses, long-term patients like stroke victims, people with extremely high medical expense needs like premature babies and things of this nature. But the one that's like a social issue you could really do something about is people who experience homelessness. And so there is definitely reason to say we should be thinking about them differently than we're thinking about just the average population because their cost drivers are particularly salient and social in nature.

And there are things that are modifiable. We can change them. A lot of things, you can't change like the fact that someone got hit by a bus and ended up in the hospital for a year and a half. Obviously you can improve traffic safety, I don't mean to belittle that. But certainly in the case of homelessness, the conditions are more modifiable and more directly targeted because we know who they are, we know where they are, we know what we should be doing.

Tristan Markle (29:16)
Yes, so in terms of the two critiques we hear of cost offset analysis, one which was the fixed cost, could be addressed by just looking with a longer time frame. But what about across departments or jurisdictions? That seems like it requires a different kind of advocacy approach. Have you seen any examples where there's cost savings in one jurisdiction because of an intervention with homelessness, where those jurisdictions were able to work together?

Dennis P Culhane (29:47)
It’s not so much cross jurisdictional as it is level of government, where the federal government has many more resources and is able to bear the burden of some costs like housing much more fairly and equitably than a local government, which just doesn't have the revenue to sustain long-term payments for rent like that, for example. So there are definitely levels of government where the trade-offs have to be thought about. You know, in most local governments, the costs they incur around homelessness are not at the individual level. They're more at the community level. It's the parks, the sanitation, you know, the fire departments and emergency services, the police, the impact on libraries and schools. There's hardly a public institution that is not impacted by homelessness, especially the unsheltered homeless population that is having to live their daily activities in public, in public view. And it has a lot of secondary effects and costs for local governments.

Tristan Markle (31:03)
Yeah, and when I was saying jurisdiction, I was mostly thinking about order of government or level of government too. And I don't know that we need more research. One could maybe just do a literature review and maybe look at your study and other studies on the cost of homelessness and just do a matrix on each order of government to show that they all will benefit to a certain degree. Jails in Canada, jails would be mostly federal. Hospitals would be mostly provincial. Policing and those other areas are municipal. So if we could maybe show the cost of homelessness across orders of government, maybe we could get them around the table. What do you think?

Dennis P Culhane (31:38)
Right. Yes, I think that's a very interesting approach and we should be doing more analysis of that kind. Because it not only shows where the costs are, but it shows where the reduced costs could be imagined. And I think, for example, we have lots of unsheltered people, let's say in California, and even if they have a disability that's been acknowledged and they get disability assistance, it's very low. $900 to $1000 a month is what they receive. It's not enough to enable them to afford an apartment. And the question is, how do we get the political will to mobilize an increase in use of state or federal dollars for rental assistance, to help people to take their basic disability income plus some rent to be able to afford a unit, whether that's in their own apartment or whether they're sharing a room in an apartment with other family members or in a boarding situation?

We need to have bipartisan support for those kinds of things. And I think that one of the interesting dynamics I envision is that you could see that, because this is really affecting everyone in California, everybody of any political persuasion knows that the homeless problem is severe and it's having a big impact. You could get the congressional delegation, it seems to me, for California, for Oregon and Washington state, some of these big states that are impacted, and probably could create a bipartisan influence and win in the US Congress in normal times. We're not in normal times right now. But I think that there's the potential for a political coalition among the most heavily impacted places to try to get the Congress to make some investments in this where it would have big benefits across the board.

Tristan Markle (33:40)
Yeah, and you did have some success in the 2000s with addressing chronic homelessness and family homelessness veterans homelessness. Can you speak a little bit about how that came about starting in the 2000s?

Dennis P Culhane (33:55)
Sure, the chronic homeless initiative was started in 2003 when George W Bush was president, and there was political pressure brought to bear from the chronic homelessness advocacy community, if you will, and the supported housing community to get the feds to try to do something. Bush's slogan was compassionate conservatism. Reducing homelessness and saving money sounded like it fit the definition of compassionate conservatism. So the president was persuaded to get behind an effort to invest about $500 million a year into supporting apartments for chronically homeless people. And about 60,000 units, apartments, were ultimately funded through this initiative between 2003 and 2008, and those units remain today. They turn over the occupancy, but they're targeted to chronically homeless people. It led to a big increase in the inventory. And even though we didn't have great data to track the population-wide numbers, we have very good reason to believe it cut chronic homelessness by about 35% between 2003 and 2008. That was big.

Then when Obama was elected and came in, the focus that the Obama administration put was on veteran homelessness. So we were able to take the same lessons we learned in chronic homelessness, and apply it to veterans with a program specifically called VASH, the Veterans Affairs Supported Housing program for chronically homeless.

Plus we created a new program for rapid rehousing and prevention of homelessness among people who were in crisis, the transitional homeless population – the much larger group that we had not focused on before. But every chronically homeless person starts out as a short-term homeless person. Our failure to actually try to address short-term homelessness was leading to chronic homelessness. We knew that. So we needed to have some kind of intervention for that population. So anyway, we built a combination of those things and veteran homelessness declined by 50%, just during the Obama administration. So over that eight year period, we saw a substantial decline in veteran homelessness. It was pretty remarkable, really. We hadn't registered something that big before. At the same time, chronic homelessness went down another 35% between 2008 and 2017. And then family homelessness went down 35% as well.

And that was mostly because we went from a policy of transitional housing, a very expensive program that paid people's rent for up to two years, with lots of intensive services, so it was a very high cost-per-unit kind of intervention. And we knew we could help many more people with a shorter shelter stay and help getting into an apartment immediately. And that's really what happened in the case of family homelessness. We moved from a transitional housing to a rapid rehousing model. And it was much cheaper and it reduced the length of time families were stuck in homelessness.

Tristan Markle (37:27)
If I could summarize, there were initiatives that you were involved in, and many more, I'm sure, across the country, where there was a housing first approach, where just generally there was more of an investment in helping people find housing sooner. And part of the argument was that it cost a lot of money to not do that, and so why not do that? And that was successful in reducing homelessness over the period of 2008 to 2015.

But then homelessness started to grow, on average, across the country, and I think in Canada too, starting in 2014. We're in Vancouver here, we saw the rents start to skyrocket in 2014, and I think it was the same in the United States...

Dennis P Culhane (38:10)
Well, in the United States really there were eight cities that experienced this huge spike. Principally they're the West Coast cities. So we're talking like Seattle, Portland, up and down in California from San Francisco down to Los Angeles and San Diego. But there was also Denver. We had New York, DC and Austin, Texas. So there really were eight places that really jump out, where you see that the rents accelerated at a very high level in that period of time and led to a huge growth in homelessness. Yes, on average, homelessness did go up across the United States, but only like one or 2% a year, even though it was going up at double digit rates in these communities we just mentioned. In order for that to happen, that means that homelessness is actually continuing to decline in most other parts of the country. And that's why the national numbers don't really show that huge explosion. It's because there's declines happening, but it's mostly geographically, the increases are geographically concentrated.

Now, since then, there is some reason to believe we're experiencing a generalized increase in housing prices across the rental market, across the country since the pandemic. And there's a number of reasons for that. I think everyone knows that the cost of the housing industry, all these things, are hitting home, and across our country we are seeing that. That gives us great reason for concern.

Tristan Markle (39:59)
You were involved in a very important paper, it was influential for me and others I know, about the community-level determinants of homelessness, or maybe in everyday language: what's causing homelessness? Looking at a whole bunch of factors, individual risk factors, or economic factors in the different cities or counties. And this is something I think is important for people to understand. Could you distinguish between an individual risk factor on the one hand and an economic factor on the other?

Dennis P Culhane (40:34)
Sure, so I did a couple studies of this nature. Some inside the VA, sheltered, unsheltered, veteran, non-veteran, also did a study with Zillow, the real estate company, and some economists there, where they had very good data on property-level prices and rents things of this nature. So the overall picture is consistent that community-level factors are the primary determinants. Community housing market factors, in particular, are the largest driver of the prevalence of homelessness. Why some communities have more homelessness than other communities have. So that's been established. But it's not the only phenomenon. There's multiple layers of life that are occurring. Yes, you have housing markets, but markets then get mediated at a neighborhood level. It gets mediated in a community where you have pressures around evictions. You have pressures for gentrification that are converting old units to newer purposes, higher cost purposes. You have people having to crowd in housing as a result of there not being enough units for them to afford. So families have to double up, or young people have to live with their parents longer than they would because they can't embark on their own getting their own place. Family formation is stalled.

So there's a ripple effect of the lack of affordable housing. And it does create these pressure domestic situations where there's a risk for some kind of fallout within the family, whether it's a disagreement about money or a disagreement about partners or behavior, relationships, whatever it may be, the stuff of everyday life. And that leads people to have to choose to leave that situation or they get forced out.

That's ultimately what happens. And there's been some great research to show that it used to be thought that individual level risk factors were the most important in terms of who becomes homeless and who does not. But then a bunch of studies came along. We call them case control studies, where you get people from the same pool of people, the same population, same neighborhoods, let's say. And you look at those people who became homeless and those people who didn't become homeless.

And when we compare them, they don't look different. They have the same background factors, the same level of these risk factors. They're not more likely to have a mental health or substance abuse problem. They're not more likely to have a criminal history or less education. They look the same. What happens is they're all pretty much at high risk and random events are striking them. And homelessness is usually… its onset is associated with a triggering event.

So you may be at risk for all these environmental factors. Your access to alternatives to get another apartment, to be able to find a place to go is very limited. But there's a triggering event that makes your housing disrupted. And that usually involves things like an accident or an injury or a hospitalization. It could involve a sudden loss of a job, a death of a spouse or family member or a wage earner in the household. The birth of a child, you know, can increase the cost of burden in a household. So there's sort of these predictable everyday life events that happen to people, happen to families, and they are the things that tip people over into homelessness. And so what distinguishes them is not their underlying risk, it's that they experience one of these triggering events.

Tristan Markle (44:18)
Reading papers on the community-level determinants of homelessness, really what it looked like is that there's some counties and cities where rents are higher and that's the biggest driver. For every $100 increase in rent, there is a 15 to 40% increase in homelessness in those areas. And that might seem like a no-brainer that increasing rents can lead to homelessness, but it didn't seem like it was necessarily a no-brainer until there was more research exploring that and backing that up.

Dennis P Culhane (44:54)
I was gonna say people have sort of two narratives about homelessness out there. One is that people say the rents are too high, there are too few places for people to go. So there's that sort of economic argument. But at the same time, there's also the people who visually you can recognize as homeless. And you say, these people haven't just become homeless because their rent went up. These are people who have experienced really challenging mental health, substance abuse problems, disabilities, et cetera. But the fact of the matter is … yeah, go ahead.

Tristan Markle (45:24)
Here's my oversimplification. After reading Homelessness is a Housing Problem, a book that's kind of inspired by your research, I think, they give an example of musical chairs where they say, okay, you have 10 chairs, and a chair is lost. Maybe the rents went too high for the chair or it was demolished or gentrified, whatever. So now you have 10 people, 9 chairs. Someone is not going to get that chair. My oversimplification would be that those factors and other community factors you mentioned determine the number of chairs, therefore the number of people who are homeless. But the question of who will not get that chair that's lost, that's the individual risk factor, the random event that happens to you. How would you respond to that simplification?

Dennis P Culhane (46:16)
I think that's fair. And really there's two causal processes going on. There's things that cause people to become homeless, and then there's a whole separate set of factors that cause people to remain homeless. And those are two separate phenomena. And so you can have a phenomenon, it's just like COVID – everyone sort of became a little epidemiologist with COVID, We all learned a lot about that – there's differences between people who get exposed and then get COVID, compared to people who have a bad outcome with COVID. The people who have a bad outcome, their exposure isn't different. What's different is that they have underlying health conditions that are going to make their care more complicated, or they get for whatever reason, a significantly different kind of dosing of this virus and have long COVID or whatever the case may be. It's completely reasonable for us to understand that there's certain factors that cause homelessness and those are different than the factors that cause people to have long-term homelessness and a bad outcome. And we need to be able to hold those two things in mind simultaneously because they are two truths that exist. By the way, with your musical chairs analogy, I always like to point out that if you're disabled and you're in a wheelchair, you never lose in musical chairs. You're always protected.

Tristan Markle (47:41)
The last bit of questions is about going a little bit beyond the cost offset discussion on the cost of homelessness, because I think it's a little bit difficult to only think in terms of the cost to services when we're thinking about the cost of homelessness. One thing you wrote in a review was that “service utilization research is limited in that it is based on a cost accounting approach to cost analysis, and more comprehensive economic studies would monetize these other aspects of homelessness, including the value to persons and to society of having stable housing, improved health, employment prospects, with family members.” Can you speak a little bit to those? And has there been any research on those in the past few years.

Dennis P Culhane (48:33)
Well, I think there has not been enough on that and that we do need to tell the fuller story of what is the great advantage of a higher quality of life in our community and to the public. And that it's not just about the people who are the recipients of particular services that benefit, we all benefit collectively from having greater security. We don't even know when we might ever need something like that or someone in our family might need it. You only learn about it when it's not there. So I think that it's something we need to put more of a spotlight on.

I would also argue there's a moral cost to our society of having people living destitute on the street. What are we telling our children about how we care as a society for the most vulnerable people, people who have disabilities, people who are elderly, who cannot really care for themselves financially? What kind of society are we that we would do that? Or other people from other countries come and visit the United States they have these real idealistic views of America a wealthy place, the richest country on earth, and yet they see that we don't care for these people who are so needy among us and we're not providing any basic protection. So I think we have a moral cost that may be more significant than all of these things. And again, it's one of those things that we know by its absence, but we really don't know how great it would be to be like, say, more Canadian, where you have more social protections. And how much happier we would be as a society, perhaps.

Tristan Markle (50:20)
Or Finnish.

Yeah. There's different ways of looking at it. Some things that we do look at are the life expectancy or the mortality rate of people who are homeless. It really just gets at the brutality of forcing people to live like that. A study where we are in British Columbia found that people who are homeless had half the life expectancy. Others show that their chance of dying is 5 to 10 times higher when they're homeless. These are starting to get at how brutal it is to force people to live like that. I did notice that in the UK, health economists are starting to look at the quality of life of homelessness and how it almost compares to being on dialysis, which is one of the worst qualities of life. Not that quantifying it in that way is presenting the whole picture. But I did notice that maybe one of your colleagues, Thomas Byrne, was involved in a paper looking at quality of life among people who were using Housing First programs (as compared to being homeless) and saw some really important results just last year. Did you look at that?

Dennis P Culhane (51:27)
Yes, the VA, Tom Byrne, all the people working there have been very proficient in producing more results that show the benefits and the cost effectiveness of a lot of these programs. I think that it's critically important information. And having a federal agency that's actually providing services, a full range of health services, social services, housing to people who are veterans gives the VA a unique picture on some of these things. So it's not surprising that the kind of research you're talking about that Tom and others are doing is coming out of the VA where they have really good data resources and they have the time and staff and energy to put into thinking about these things. There's something to be said as well for just having infrastructure of researchers and thinkers who can focus on these things, and the value that that brings to what we do in the future.

Tristan Markle (52:39)
Yeah, and last question, when we set up this interview, it was before some results in Washington, DC, some reporting recently about Housing First. Can you let us know what's happening with Housing First and what can we do about it?

Dennis P Culhane (52:56)
I think there's really two minds going on here, interestingly. On the one hand, we are seeing that the proposals, the Trump budget proposals, are to scale back subsidized housing dramatically and change the way it's paid for in the United States, including a $500 million cut in homeless-specific housing funds. That would be pretty devastating. Part of that is also to say we no longer will accept only housing first programs, but we want other kinds of programs prioritized as well. Mostly those are various forms of transitional housing, which, like I said, we did that for 20 years from the mid 80s to the mid 2000s, and it was the failure of those programs that led us to housing first. But in any event, there are people who want to go back and do that. So we have that on the one hand, kind of a very not so subtle and pretty right upfront sort of attack on where we have been.

On the other hand, with the veterans program, the President's budget is a $1.1 billion increase, which would be almost a 50 % increase in spending on veteran homelessness. And they're talking about, not just sustaining the people in housing that they have, but how are we going to find other housing options for the people who remain? Relatively few. We have about 30,000 veterans who are homeless on a given night right now, down from 90,000 or so. So we've made a lot of progress. No one’s talking about getting rid of housing options for the people who are veterans who are in need. They're still talking about what are the new housing options we need? What are the things we need to do?

So, on the one hand, we have, “there are veterans, they're homeless, we have to do something about it”. And the people who are not veterans and are homeless, we have to cut the funding available to them. It doesn't make sense. It's contradictory. That's the moment we're in.

Tristan Markle (54:56)
Well thanks for the update on that. Hopefully we can all do our part to make governments design homelessness policy in a way that is consistent and based on evidence.

Dennis P Culhane (55:09)
Well, thank you. Hope that your podcast helps get the story out.

Tristan Markle (55:10)
Thanks so much, Dennis. Thanks for inspiring a whole generation of researchers and helping to inspire our podcast series.

Dennis P Culhane (55:15)
You're welcome. Great. Well, thank you. I look forward to hearing more about it.