Stop Giving Out Tents, Start Giving Out Wood

This is Part 2 of a series. Part 1: Why I Had to Learn About Homelessness explains how I got here.

Last week I wrote about how conversations with taxi drivers led me down a rabbit hole on homelessness. The short version: I kept losing people on climate because they’d point to homelessness as proof that government can’t solve big problems. So I went and learned about homelessness.

What I found: we have the intervention order backwards. We try to stabilize people, then treat them, then restore their dignity. It should be the reverse. Dignity first, then stability, then treatment if they still need it.

And the single biggest barrier to dignity? Tents. The equipment of transience.

Today I want to get into the numbers.

The Numbers Are Real. And So Is the Failure.

Here’s the truth: homelessness has increased. The rate has roughly doubled since 1990.

Period	Rate (per 100,000)	Context
Great Depression (~1933)	~1,200	Peak crisis, Hoovervilles
1984	85-213	Federal estimate range
1990	~92	US Census
2019	171	HUD/USICH
2024	~230	HUD estimate

Sources: US Census Bureau, HUD Annual Homeless Assessment Reports. Note: These figures come from different methodologies (Depression-era estimates, early federal ranges, Census attempts, modern HUD point-in-time counts) and aren’t directly comparable. The direction is clear; the precision isn’t.

That’s a real increase. I’m not going to pretend otherwise.

But here’s what I keep coming back to: everything we’ve tried hasn’t worked. California spent over $20 billion on homelessness programs between 2018 and 2024. (This includes healthcare, mental health spending, and housing programs combined, per state audit reports.) The result? Homelessness increased by 6%. Los Angeles passed a $1.2 billion bond to build supportive housing. They promised 10,000 units at $140,000 each. By 2022, they’d completed about 1,200 units at $596,000 each. 325% over budget. Four years behind schedule.

We don’t have a funding problem. We have a direction problem.

Who This Works For

Who am I talking about?

This approach works best for the situationally homeless: people who lost housing due to job loss, eviction, domestic violence, or medical crisis. The working homeless who have jobs but can’t afford first/last/deposit. Families caught between paychecks and rent increases. Veterans who fell through the cracks. People who are capable but stuck.

This isn’t a replacement for intensive psychiatric services for the chronically mentally ill. It’s a way to stop wasting those scarce resources on people who just need stability and a key. When you give everyone the same intake process, the same case manager caseload, the same shelter bed, you’re treating a recently evicted nurse the same as someone with severe untreated schizophrenia. That helps no one.

Separate the populations. Give the capable a path back. Then focus intensive resources on those who actually need them.

What Austin Understood

Community First! Village in Austin, Texas is the proof of concept.

It’s a 51-acre development that currently houses 393 formerly homeless people, with plans to expand to 1,900 homes. The average cost per tiny home? About $30,000.

Now, I need to be honest about what that number does and doesn’t include. Austin’s $30,000 covers the structure itself. It doesn’t include land (the 51 acres was donated). It doesn’t include the ongoing operational costs of running the community. And tiny homes are a different product than permanent supportive housing: they’re smaller, simpler structures without the same level of built-in services.

Compare that to LA’s $596,000 per unit, which includes permanent construction to code, wraparound services, and land purchased at Los Angeles market rates. These are different interventions with different cost structures and different outputs. The national median for permanent supportive housing is closer to $250,000-$350,000 per unit. LA and San Francisco are high-cost outliers.

What the comparison does show: there’s a massive gap between what we’re spending and what we’re getting. And there are other models worth understanding.

Here’s what makes Austin’s model interesting:

100% privately funded. No government grants. Pure charity, funded by donations and run by Mobile Loaves & Fishes, a Christian nonprofit with a perfect 100/100 score on Charity Navigator.
Residents pay rent. About $300 per month. If you don’t pay, you don’t stay. This isn’t a handout. It’s housing with expectations.
On-site employment. An auto shop where residents can earn up to $1,500 per month. An art studio. A woodshop. A forge. A farm. Dignified work, not just a bed.
Community accountability. Neighbors mediate conflicts. There are rules. There are consequences.

The philosophy that drives it: “Housing alone will never solve homelessness, but community will.”

I’m not claiming Austin has solved homelessness. They have their own challenges, their own selection criteria, their own context. Replicating this model elsewhere would require either similar land donations or public land allocation. And LA’s problems are completely different. If anyone had a scalable solution that actually worked, we would all know about it by now.

But Austin is trying something worth understanding. The tiny homes aren’t the point. The dignity is.

The Disaster We Won’t Name

Here’s something that bothers me.

When a hurricane hits Florida, we don’t hesitate. Volunteers stream in. Habitat for Humanity coordinates build days. Churches send mission teams. Communities rally to rebuild homes for people who lost everything.

Nobody asks whether the hurricane victims made bad choices. Nobody points out that they chose to live in a hurricane zone. Nobody suggests that helping them creates moral hazard.

We just help. Because they’re facing a disaster.

Homelessness is also a disaster. Someone loses their job. Their car breaks down. They get sick and can’t afford care. Rent goes up. One thing leads to another, and suddenly they’re sleeping in their car, then on the street.

I know this because I lived a version of it. I fell off a cliff. I needed medical care I couldn’t afford. The system offered me permanent dependency or nothing. I spent years fighting my way back, and I never forgot what it felt like to have no path forward.

Why is homelessness a less worthy disaster than a hurricane? Why do we mobilize for one and not the other?

Habitat for Humanity has built over 800,000 homes worldwide using community volunteers and sweat equity from the homeowners themselves. We know how to rally together. We know how to build. We do it all the time for disasters we’ve decided are worthy of our compassion.

Homelessness is a disaster. We should treat it like one.

The Evidence for Stability

The research is clear on one thing: when people have a stable place to be, services work. When they’re scattered and swept, services fail.

Transition to permanent housing:

Tiny home villages: 15-40% of residents transition to permanent housing (Seattle’s LIHI Village Network, 2019-2022; Portland’s village programs report similar ranges)
Dispersed encampments: 3-7% (HUD Street Outreach Program data, 2021)

Service engagement:

Stable settings: 70-85% engage with case management (RAND Corporation, “LA LEADS Evaluation,” 2022)
Street populations: less than 30% consistently engage (same study)
Mobile populations miss 60% of scheduled appointments (HUD Office of Policy Development and Research, 2020)

Case management outcomes:

Clients in stable housing are 2.3x more likely to complete housing navigation programs (Urban Institute, “Stabilization Services in Permanent Supportive Housing,” 2020)
Street outreach averages 18 months to housing placement vs. 6-9 months from a village setting (National Alliance to End Homelessness, “State of Homelessness” reports, 2019-2022)
Case managers can effectively serve 15-20 clients in stable settings vs. 8-10 when tracking mobile populations (Corporation for Supportive Housing, caseload guidelines)

Health outcomes:

Unsheltered individuals average 4.2 ER visits per year vs. 2.1 for those in stable housing (Boston Health Care for the Homeless Program, Annual Report 2021)

The pattern is consistent across studies: stability enables services. Dispersion defeats them.

The Mental Health Question

I need to address this directly.

We have a mental health crisis in this country. We’ve outsourced care to institutions that are overwhelmed, underfunded, and often ineffective. And here’s the uncomfortable truth: there has never been a society that handles severe, untreatable psychiatric disorders in a way we universally applaud. There may be no right answer yet.

But we know we’re doing the wrong thing. We all know it.

Here’s what I propose: give people dignity first. Give them materials. Let them build. Let them form community. Let them take care of each other.

Some will take the road back into mainstream housing. Some will need ongoing support. That’s fine.

But then we’ll actually know who needs intensive help. Not because they failed a screening administered by a stranger, but because they had every opportunity to rebuild and still couldn’t. We will be able to see who the actually sick are once we’ve given everyone else their dignity back and a road back into society.

What emerges is a clearer picture of who needs intensive, long-term care. Not a mass of undifferentiated suffering, but specific people we can actually see, count, and help. Instead of the chaos we have now, where the recently evicted family and the person with chronic schizophrenia are both sleeping in tents, both getting swept, both invisible.

Give everyone the basics first. Most will stabilize. The ones who can’t will become visible, not as failures, but as people who need a different kind of help. Then we can actually focus our intensive resources where they’re needed.

The Cost Comparison

Approach	Cost Per Unit	Build Time	Notes
LA Permanent Supportive Housing	$596,000	3-5 years	High-cost outlier; includes land, services
SF Permanent Supportive Housing	$700,000-$1.2M	3-5 years	Highest costs nationally
National PSH Median	$250,000-$350,000	2-4 years	Varies by market
Tiny Home Village	$15,000-$65,000	2-6 months	Structure only; land/ops separate
Community First! Village (Austin)	~$30,000	Months	Land donated; 100% private
Pallet Shelters	$6,500-$15,000	Days	Immediate stability

California spent $20 billion and homelessness increased. Austin is trying something different at a fraction of the cost, though with donated land and a different model. Neither has solved the problem. If we had a scalable solution that worked, we’d all know about it.

What This Is and Isn’t

Here’s what I’m proposing.

This is:

Managed villages with structure, screening, and accountability
Designated sites on underutilized urban land (industrial zones, city parcels)
A pathway for the situationally homeless to rebuild stability
An alternative to the tent-and-sweep cycle that helps no one

This is not:

Dumping people on a lot with no support
Lawless encampments with no rules
A replacement for mental health services
Suburban subdivisions with villages next to your cul-de-sac

The villages that work have governance. Austin has a community council. Portland’s villages have elected leadership. There are rules about violence, about paying rent, about maintaining your space. There are consequences for breaking them.

One thing I won’t hedge on: anyone who hurts a child or is violent goes to jail. Full stop. Society falls apart without the protection of children and the elderly. We already have a place for people who can’t follow that rule, and it doesn’t require a caseworker to enforce it. That’s what the justice system is for.

Who Pays and Who Decides

Let me address something I’ve glossed over: land and money.

Most successful models use donated private land (Austin) or underutilized public parcels. Property rights matter. Owners should be free to permit settlements on their land if they choose. Churches already do this. Ranchers do this. It’s currently illegal in most places, which is part of the problem.

And I should say this directly: site selection matters, and communities deserve genuine input. I’ve studied how this works. My graduate research examined community response to wind energy siting across 53 proposals in the Western US. What we found: opposition isn’t irrational NIMBYism. It’s shaped by whether communities have genuine processes for participation. When they do, resistance is manageable. When they don’t, it escalates. If you want villages that work, you need neighbors who had a real voice in where they go. Not theatrical hearings. Real input.

For public land, cities should identify surplus parcels, unused industrial sites, land that’s been sitting empty for decades. This isn’t about condemning private property. It’s about using what we already have.

As for who pays for materials: Austin proves this can be 100% privately funded through charity. That’s not the only model. Some cities might fund it directly. Some might partner with nonprofits. Some might just remove the legal barriers and let communities solve it themselves.

The Legal Reality

I should also acknowledge the boring but essential part: liability.

Cities that officially sanction informal construction assume risk. One fire death, one assault, one injury, and there are lawsuits. This is why most cities don’t try anything creative. The legal exposure is real.

Any real implementation needs a framework: emergency declarations that provide liability protection, new state legislation enabling managed villages, or partnership structures where nonprofits assume operational responsibility. That’s the unglamorous work that makes bold ideas possible.

I don’t have the legal answers. But I know they exist, because Austin exists, and Portland’s villages exist, and cities find ways to do things when they actually want to do them.

An Invitation

I’m not naive. I know that “just let people build” isn’t a complete policy platform. Building codes exist for good reasons, and meeting those standards costs money.

And I want to be clear about what this is: a thought piece, not legislation. My job is to point in a direction. The operators know things I don’t.

So here’s my invitation: if you work for a homeless services nonprofit, a housing authority, a city planning department, a faith-based organization, or a community development corporation, I want you to imagine you were drafting the rules for this. What would the screening look like? What would the governance structure be? What services would you co-locate? What would the rent structure look like for people with zero income? How would you handle the hard cases?

The people who have those answers are already working in this space, waiting for permission to try something different.

This is a solution for cities, not suburbs. People experiencing homelessness need to be close to services, close to stores, close to jobs, close to transit. The underutilized parcels I’m talking about are in urban cores and industrial zones, not residential subdivisions.

No one has data showing homeless fixes that work at scale. The evaluation research I cite throughout this piece (Urban Institute, RAND, HUD audits) are post-hoc assessments of what actually happened. But the consultant scopes of work that promise interventions will achieve specific outcomes before they’re tested? That’s theater. I’ve written both evaluations and proposals. I know the difference.

So here’s what I propose instead: dignity first. Give people materials. Let them build. Let them form community. Let them take care of each other. Give them the road back into society and see who takes it.

Maybe it’s time to try something very old. Something that worked for thousands of years before we decided we knew better.

First, give the people plywood. Give the people tin. Give the people piping, and let dignity back in.

(Yes, I know. The rhyme is corny. I’m keeping it anyway. Sometimes the memorable thing is more useful than the sophisticated thing.)

Questions I Expect to Be Asked

“Aren’t you romanticizing poverty?”

No. I’m romanticizing mainstreet. I grew up in a small town. I’ve lived in Seoul, where hundreds of vibrant main streets keep disabled, elderly, young, rich, and poor all together in the same neighborhoods. I want that back.

“What about fire safety and building codes?”

Building codes increase costs, and that’s protection we’ve collectively agreed to pay for. I’m not suggesting we abandon safety standards. I’m suggesting we create designated spaces where people can build with basic support and oversight rather than scattering them where they build without any support at all. Every tent camp today has zero fire safety. A designated site with basic oversight, fire extinguishers, and spacing requirements is safer than what exists now, not more dangerous.

“This would destroy property values.”

You know what’s already destroying property values? Tents on every sidewalk downtown. The question isn’t whether homeless people exist near your property. They already do. The question is whether they’re scattered and chaotic or settled and stable.

“Who would actually support this politically?”

Business owners tired of sweeps that don’t work. Conservatives who believe in self-reliance and community solutions over government programs. Faith communities who already run successful models like Austin’s. Progressives who recognize that $596,000 per unit isn’t scaling. Anyone who’s honest about the fact that what we’re doing now has failed.

“So your actual policy recommendation is to sweep everyone into one area?”

No. The opposite. Right now we sweep people away from everywhere, into nowhere. I’m proposing we create somewhere they can stay.

The difference between a sweep and an invitation matters. A sweep says “you can’t be here.” A designated village says “you can be there.” One is displacement. The other is placement.

Anna Kelly is the author of Climate Champions and works on energy policy and building performance standards. Her graduate research on community response to wind energy siting was published in Energy Research & Social Science. She writes about climate, housing, and the systems that shape how we live.

← Read Part 1: Why I Had to Learn About Homelessness