Portland State University 2019 Homelessness report

From HousingWiki
cover of report

[originally part of article Right to housing . See also Reading List: Homelessness section]. 

Portland State University Homelessness Research & Action Collaborative (HRAC) on August 21, 2019 issued a report, "Governance, Costs, and Revenue Raising to Address and Prevent Homelessness in the Portland Tri-County Region." [Zapata et al 2019]. 

"HRAC's goal is to help reduce homelessness and its negative impacts on individuals, families and communities. The Collaborative brings together the expertise and skills of each of Portland State University's colleges and schools, and collaborates with people experiencing homelessness, advocates, service providers, city and county policymakers and other stakeholders." 

HRAC Faculty & Staff: 

  • Marisa Zapata. Center Director - Urban Studies & Planning
  • Greg Townley, Director of Research. Psychology Department
  • Jacen Greene, School of Business
  • Lisa Hawash, School of Social Work
  • Todd Ferry, School of Architecture
  • Maude Hines, English Department
  • Sergio Palleroni, School Architecture
  • Paula Carder, OHSU-PSU School of Public Health
     

Summary / analysis 

Introduction:

"This report takes a comprehensive look at the scale of homelessness and housing insecurity experienced in the Portland tri-county area. Our goal in producing this report is to help community members understand the scope and scale of the challenges we face when addressing homelessness and housing insecurity. We examine governance options, provide cost estimates for providing housing, supports, and services, and present revenue-raising options for our local governments to address homelessness and housing insecurity."

Expanded definition & scope of 'homelessness' and precariously/unsuitably housed 

"Many of the available counts of those experiencing homelessness use a narrow definition. We believe this leaves people behind. For example, the official Point-in-Time counts do not include those living doubled up, those sometimes described as the hidden homeless or precariously housed. This vulnerable population is sleeping on friends’ couches or cramming in unsafe numbers into bedrooms."

Population and cost estimates

"38,000 people experienced homelessness in the tri-county area in 2017...The cost to house and support this population ranges from $2.6 billion to $4.1 billion over ten years based on a range of options presented in the cost section of this report."

for reference, the Metro and Portland housing bonds passed in the last two years total approximately $911 million for affordable housing. 

"As many as 107,000 households faced housing insecurity or were at risk of homelessness in 2017 in the tri-county area due to low incomes and paying more than 30% of their income on housing costs....Providing rent assistance for all of these households would help resolve housing insecurity and reduce the risk of becoming homeless. We estimated costs to create such a program, using a range of rents and addressing households that earn 0–80% of the median family income (MFI) for their household size. To help severely cost-burdened households [defined by HUD as paying more than 50 percent of one's income on rent. -YIMBYwiki] over ten years would cost $8.7 billion–$16.6 billion. That’s about $870 million–$1.66 billion per year, or $10,000–$20,000 per household per year. These numbers do not account for what is already being spent in the tri-county area to relieve the cost burden for households in need."

 

Source of cost estimates:

"The costs of developing housing units, including new construction and rehabilitation, mainly follow the vetted assumptions from the Corporation for Supportive Housing (2018 and 2019) reports (based on “actual costs reported by PHB and approved by stakeholder advisory groups”). The only adjustment comes from the Metro Affordable Housing Bond Program Work Plan (2019) and Regional Housing Bond Financial Modeling Summary Memorandum (2018). These sources peg the average construction cost of housing units at $215,000 (a weighted average for all housing unit sizes), and the cost of rehabilitation of existing units at $190,000 (including $150,000 building acquisition cost and $40,000 rehabilitation cost, all in 2018 dollars). CSH (2018) estimates that annual operating and maintenance costs run between $6,000 and $8,000 per unit."

 

Related comments from study authors 

Zapata, Marisa. "My View: Homelessness isn't the crisis you think it is. [Homelessness is a warning that our systems are failing. How we rebuild is up to us]."  Portland Tribune, Opinion, August 15, 2019

"the one true solution: stable, quality, affordable and supportive housing." 

"To truly stop homelessness, we need to fix the long-term systemic issues that lead to it in the first place. We need universal rights to safe and quality housing and universal health care so no one is forced to choose between buying insulin and making rent. We need major criminal justice reform that does not sentence people to life on the streets and an overhaul of how we support those with addiction and major mental illnesses so recovery is possible for everyone. We need stronger protections for historically marginalized communities so your ZIP code, gender identity or skin color doesn't dictate your future."

 

{{#widget:Tweet |id=1166039336644751366 |conversation=none }}

{{#widget:Tweet |id=1163862980187975680 |conversation=none }}

{{#widget:Tweet |id=1167105482646904832 |conversation=none }}

 

Right to Housing aspect

This report does not explicitly articulate a "Right to Housing" concept, but it notably uses broad definitions and makes wide estimates of how many people in the region experience homelessness in some form, or are considered vulnerable to falling into homelessness. By proposing government assistance, in the form of housing provision and rent assistance, for these entire populations, it in effect approaches an idea of "Right_to_housing." 

"This report takes a comprehensive look at the scale of homelessness and housing insecurity experienced in the Portland tri-county area. Our goal in producing this report is to help community members understand the scope and scale of the challenges we face when addressing homelessness and housing insecurity. We examine governance options, provide cost estimates for providing housing, supports, and services, and present revenue-raising options for our local governments to address homelessness and housing insecurity.

Before getting too far into the report, we want to make sure to note a few things. Many of the available counts of those experiencing homelessness use a narrow definition. We believe this leaves people behind. For example, the official Point-in-Time counts do not include those living doubled up, those sometimes described as the hidden homeless or precariously housed. This vulnerable population is sleeping on friends’ couches or cramming in unsafe numbers into bedrooms..."  

 

 

Public responses to report

{{#widget:Tweet |id=1165727139552886789 |conversation=none }}

 

Questions / critique

Why use static, not dynamic figure for homeless population?

What are the implications of basing the cost estimates on a static figure for current homeless population, rather than current population plus observed or estimated typical rate of 'inflow' to homelessness? 

The report states: 

"The type of modeling needed to capture the inflow and outflow of people experiencing homelessness is complex, data intensive, and time consuming. We opted to go in the opposite direction, and created replicable, straightforward estimates completed in just a few months. Our goal was to provide a general sense of the number of households and associated costs, and we believe that adding layers of complexity where assumptions are added to assumptions would not get us to a better estimate."

Is that true? It seems there are simple estimates routinely given for rates of inflow into homelessness, based for example on considering counts of people who were placed into housing, versus subsequent Point in Time counts. In any case, we know this inflow is substantial, certainly nonzero, and so isn't any simple nonzero estimate a better estimate than zero?  The consequence of assuming zero inflow is a likely underestimate of response costs, perhaps to a major degree. 

In addition, in debates over homelessness, it is common for parties of all types to consider that there is steady inflow of people needing services; and potentially, that provision of services could motivate or facilitate greater use of those services, or an increasing population identifying / identified as in need of them. In fact, Portland / Multnomah County, and its Joint Office of Homeless Services, experienced a well-publicized apparent case of this with its family homeless shelter program in 2017-18: 

"The emergency shelter run by the nonprofit Human Solutions...funded by the city of Portland and Multnomah County through the Joint Office of Homeless Services, had a policy that was unique on the West Coast. It promised to provide a bed for every child that needed one. It didn’t turn any family away and had no wait list.  
     And on nights when all 130 beds were in use and the other roughly 50 beds dedicated to homeless families in Portland had people sleeping in them, the county paid for hotel rooms for the extra families.
     This no-turn-away policy was a point of pride for Multnomah County, which was working to expand shelter access in response to Portland’s housing crisis.
     When the number of homeless families sleeping on the street dropped in the county’s count of homeless people in 2017, it credited the no-turn-away shelter with making the difference. 
     Just a handful of other cities — including New York and Washington, D.C. — have similar policies. Massachusetts is the only state that has recognized a right to shelter for families.
     But while Multnomah County celebrated its no-turn-away policy, it wasn’t prepared to deal with the financial and practical challenges of managing it. 
     By late 2017, the no-turn-away policy had plunged Portland’s family shelter system into financial crisis because the number of families that needed a place to sleep kept rising.
     In February 2016, when the shelter first opened, it served about 150 people a night using the shelter and overflow space in a church across the street.
     That fall, the number of families at the shelter and the overflow space climbed to around 200 a night, and staff with the Joint Office of Homeless Services grew concerned that the shelter was dangerously overcrowded.
     So, the office authorized the shelter to provide motel vouchers to homeless families when the shelter filled. 
     Last June, the number of families in the system began to rise again. Over the summer, it doubled.
     By October 2017, the nightly census hit 468 people. The number of families staying in motels, paid for by the city and county, outnumbered the families at the shelter itself.
     'It frankly put our system into crisis. It became unsustainable for us,' said Mark Jolin, director of the Joint Office of Homeless Services, which coordinates shelter for Multnomah County and the city of Portland."  [Templeton 2018].

 

Why use static, not dynamic figure for housing-vulnerable population?

Likewise, how valid is it use a static figure for housing-vulnerable (107k people?) who might receive rental assistance? Could we make some simple/reasonable prediction of rate of inflow into this group, and include that as a baseline, rather than baselining on zero inflow? 

 

What about induced behavior, ie how proposed policies would change behaviors? 

As Ellen & O'Flaherty [2010] note in the Introduction to How To House the Homeless [2010], policies alter the environments they address: 

"This volume explores a middle ground between two opposing views of how housing markets affect homelessness...one view implies that the question we are trying to answer is impossible to answer (housing market impotence); the other, that it’s trivial (housing market omnipotence)....Housing market omnipotence, looks at homelessness simply as the difference between the number of available homes and the number of households seeking housing; the only way to resolve this shortage is to build more homes and put homeless people in them. But housing markets are much more complex than simple counts.....Housing market players adjust to changing environments by moving in with relatives or moving out; by building, renovating,
or abandoning, [etc etc].
     "In the end the final outcome of a policy’s implementation may look nothing like its initial thrust or intended goals." 

 

How accurately/efficiently could those most at risk of homelessness be targetted? 

Ellen & O'Flaherty [2010b] notes:  

"Olsen argues that we should focus housing assistance on the poorest households, because these households are more likely to become homeless. However, predicting in advance who will become homeless is extremely difficult (again, O’Flaherty gives a theoretical reason for this empirical regularity), and targeting more precisely is therefore very difficult. "Third, moral hazard inevitably comes with better targeting. You can target services precisely if you are willing to accept a great deal of moral hazard; you can reduce moral hazard a great deal if you are willing to have poor targeting. But you cannot have both precise targeting and low moral hazard. If only people who are actually homeless receive great services, too many people will be homeless. If everyone receives services, only a tiny percentage of the benefit will inure to people who would otherwise be homeless."

The PSU study asserts that homelessness could generally be prevented by providing full housing assistance (e.g. to let households pay no more than 30% of income on rent) to all cost-burdened, low-income households. But what data or evidence do we have to evaluate what portion of those who fall into homelessness come from these various levels of of cost-burdened households? Or that the rental assistance would prevent their shift into homelessness? For example, a major precursor to female and family homelessness is domestic violence, and this is not necessarily addressed by the head of household receiving rental assistance.  

Perhaps it would be more effective to develop better models of identifying and targeting the households most likely to shift into homelessness, and focusing assistance pre-emptively there?  (cf Destination: Home program doing that in Santa Clara County). Or, to emphasive funding rapid rehousing for those who do become homeless?

 

How much might widely-expanded rental assistance raise rents generally?

What are the implications and credibility of assuming / modeling that large-scale rent assistance, given the Portland area's supply constraints or supply inelasticity.

A familiar idea in housing research is that expanded rental benefits in supply-inelastic markets may largely cause overall increased rents and the subsidy to be in effect captured by landlords. In a recent interview on Vox, Jennie Schuetz of the Brookings Institute argues this would apply to Greater Portland: 

"To help low-income families, we should make housing assistance an entitlement — like SNAP or Medicaid — that’s available to every family that meets the income eligibility standards.
We probably shouldn’t tie housing assistance to local housing costs, because high local housing costs reflect housing scarcity, which means extra subsidy will be captured by landlords. Instead, we should tackle the shortage.
In California, the Northeast Corridor, Greater Seattle, Greater Portland, and, to a lesser extent, Greater Denver and many college towns, there is simply not enough housing being built to meet the demand to live in these areas, creating problems that no amount of subsidy or rent control can really solve."

The PSU report acknoweldges this issue on p.78:

"Another significant element not addressed by this report is the impact that providing housing assistance at a previously unprecedented level would have on the housing market. Obviously, a massive influx of government assistance into the rental market would have dynamic implications for pricing and supply. It is not possible at this stage to determine those impacts, and this report therefore takes a static approach to market analysis and assumes no change, rather than assuming an uncertain level of change."

 

How much might new publicly funded/leased housing crowd out other housing development/use? 

What are the implications and credibility of assuming / modeling that lthat large-scale new public-funded housing development wouldn't crowd out housing that would otherwise occur, creating a price-raising factor?

"A reputable body of empirical literature, for instance, shows that construction of subsidized housing is offset, at least in part, by reductions in the unsubsidized housing stock (see, for example, Sinai and Waldfogel 2005)." [Ellen & 0'Flaherty 2010b]

 

Why add new housing only by current conventional building or leasing?

Why would or should we assume that housing creation to address homelessness at large scale should be done all with current conventional housing forms? 

The report states: 

The numbers provide a starting point for conversations on the resources necessary to tackle this issue in the tri-county area, and how we might go about raising the revenue to do so.

But how are the numbers meaningful is there is not even an examination of potential ways to fulfill the goals of the proposed, greatly expanded government service program?  

For example, Portland-based Meyer Memorial Trust recently conducted a mult-year Cost Efficiencies program, identifying and funding a variety of alternative approaches to create regulated affordable housing. (see Meyer Memorial Trust [2015], and Parkhurst [2018]). Some of the projects highlighted deliver or aim at much lower unit costs than conventional development. 

More broadly, in the on-going web/book project "Village Buildings: patterns for affordable housing from Oregon," McCormick surveys a wide variety of possible approaches that have been or might be taken for lower-cost housing in Oregon. Some of these potentially offer supportive or regulated-affordable housing at much lower costs per unit than the unit costs cited in PSU study; for example 

  • 3. Accessory Dwellings - self-, startup-, or public-financed.
  • 4. Pocket Neighborhoods, Cohousing, Villages: Ross Chapin, Cully Grove, SquareOne Villages.
  • 5. Private affordable housing: built without subsidy. Guerrilla Development, Rob Justus / Home First.
  • 7. Interim/redeployable housing: e.g. PAD Initiative. 
     

Don't policies driving overall housing cost levels need to be part of the policy advocated?

To what degree do the estimates of housing needs and remedy costs follow from general area housing prices, which are strongly affected by economic climate and general rate/type of housing creation? If this factor is large, then it seems policy recommendations concerning this general housing creation would be an integral part of this report/project's scope. 

As Schuetz noted above: 

In California, the Northeast Corridor, Greater Seattle, Greater Portland, and, to a lesser extent, Greater Denver and many college towns, there is simply not enough housing being built to meet the demand to live in these areas, creating problems that no amount of subsidy or rent control can really solve."

 

 

Summary of issues

[from comment in Facebook group PDX YIMBY, 6 Sept 2019, by McCormick]

1. The PSU HRAC 2019 study presents estimated costs to "address" homelessness, but chooses not to include ongoing inflow of people into homelessness or housing insecurity -- which in many places is known to be consistently large, and (e.g. in Bay Area greater than the rate at which the unhoused are being housed). 
  The study also sets aside the cost-raising effect of widely expanding rent assistance in a supply-constrained market, and that of new subsidized housing crowding out other housing. Together, these assumptions likely mean the report greatly understates the scope of need, possibly to the point where the proposed remedies would nowhere near achieve the goal. 

2. Other key assumptions, such as that a low-income rental benefit would generally stop people falling into homelessness, seem to be asserted with quite little support. Providing full rent assistance to all cost-burdened low-income renters may be a good goal, but it may also be a very costly and poorly targetted way to prevent homelessness. Also, due to cost-increasing effects on overall rent levels, it may be a fairly ineffective way even to address housing affordability. 

3. The report has a major, essentially unstated assumption, that new housing created to address homelessness should be done in the same way and at same costs as current general market or tax-credit-financed housing. Particularly given the reasons noted above that report may far understate need, it seems to me a surprising assumption, that a hugely expanded program of public spending would not consider alternate approaches to achieve its goal.  
   Also, Oregon and Portland are quite notable for their history and broad community of people exploring alternative housing, such as the Meyer Memorial Trust's recent multi-year Cost Efficiencies study and pilot funding program. It would seem a great missed opportunity not to bring all this knowledge and built precedent to bear on the problem. 

In general, I fear repeating the typical pattern of housing assistance in many times & places: programs/ funding that help a portion of the needy, at high aspiration and cost, often to traditional middle or upper-middle-class norms; but leave out many more people, and perhaps don't even keep up with worsening need. Alternatively, I'd suggest, we might think of "right to housing" approaches, that create effective minimum standards that must be extended to all -- and that compel the [country, state, city..] to find how to achieve it for all.

 

See also

References 


 

Retrieved from "https://housing.wiki/wiki/Portland_State_University_2019_Homelessness_report?oldid=4841"